modeling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Machine Learning\n",
    "\n",
    "In this file, instructions how to approach the challenge can be found."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We are going to work on different types of Machine Learning problems:\n",
    "\n",
    "- **Regression Problem**: The goal is to predict delay of flights.\n",
    "- **(Stretch) Multiclass Classification**: If the plane was delayed, we will predict what type of delay it is (will be).\n",
    "- **(Stretch) Binary Classification**: The goal is to predict if the flight will be cancelled."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Main Task: Regression Problem"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The target variable is **ARR_DELAY**. We need to be careful which columns to use and which don't. For example, DEP_DELAY is going to be the perfect predictor, but we can't use it because in real-life scenario, we want to predict the delay before the flight takes of --> We can use average delay from earlier days but not the one from the actual flight we predict.  \n",
    "\n",
    "For example, variables **CARRIER_DELAY, WEATHER_DELAY, NAS_DELAY, SECURITY_DELAY, LATE_AIRCRAFT_DELAY** shouldn't be used directly as predictors as well. However, we can create various transformations from earlier values.\n",
    "\n",
    "We will be evaluating your models by predicting the ARR_DELAY for all flights **1 week in advance**."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Feature engineering will play a crucial role in this problems. We have only very little attributes so we need to create some features that will have some predictive power.\n",
    "\n",
    "- weather: we can use some weather API to look for the weather in time of the scheduled departure and scheduled arrival.\n",
    "- statistics (avg, mean, median, std, min, max...): we can take a look at previous delays and compute descriptive statistics\n",
    "- airports encoding: we need to think about what to do with the airports and other categorical variables\n",
    "- time of the day: the delay probably depends on the airport traffic which varies during the day.\n",
    "- airport traffic\n",
    "- unsupervised learning as feature engineering?\n",
    "- **what are the additional options?**: Think about what we could do more to improve the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 717,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing libraries\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_theme(style='darkgrid', context='talk')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 718,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fl_date</th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>branded_code_share</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>...</th>\n",
       "      <th>distance</th>\n",
       "      <th>carrier_delay</th>\n",
       "      <th>weather_delay</th>\n",
       "      <th>nas_delay</th>\n",
       "      <th>security_delay</th>\n",
       "      <th>late_aircraft_delay</th>\n",
       "      <th>first_dep_time</th>\n",
       "      <th>total_add_gtime</th>\n",
       "      <th>longest_add_gtime</th>\n",
       "      <th>no_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2019-11-13</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA_CODESHARE</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>...</td>\n",
       "      <td>228</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2019-12-08</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>...</td>\n",
       "      <td>2576</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>UA</td>\n",
       "      <td>UA_CODESHARE</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>...</td>\n",
       "      <td>151</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-02-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>...</td>\n",
       "      <td>430</td>\n",
       "      <td>48.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019-05-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>...</td>\n",
       "      <td>977</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2019-06-12</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA_CODESHARE</td>\n",
       "      <td>AA</td>\n",
       "      <td>3696</td>\n",
       "      <td>MQ</td>\n",
       "      <td>N667GB</td>\n",
       "      <td>3696</td>\n",
       "      <td>13930</td>\n",
       "      <td>ORD</td>\n",
       "      <td>...</td>\n",
       "      <td>135</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2018-06-24</td>\n",
       "      <td>DL</td>\n",
       "      <td>DL</td>\n",
       "      <td>DL</td>\n",
       "      <td>1180</td>\n",
       "      <td>DL</td>\n",
       "      <td>N909DA</td>\n",
       "      <td>1180</td>\n",
       "      <td>11193</td>\n",
       "      <td>CVG</td>\n",
       "      <td>...</td>\n",
       "      <td>1069</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2019-10-18</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>2277</td>\n",
       "      <td>WN</td>\n",
       "      <td>N209WN</td>\n",
       "      <td>2277</td>\n",
       "      <td>13296</td>\n",
       "      <td>MHT</td>\n",
       "      <td>...</td>\n",
       "      <td>377</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2019-03-30</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>711</td>\n",
       "      <td>AA</td>\n",
       "      <td>N818AW</td>\n",
       "      <td>711</td>\n",
       "      <td>12266</td>\n",
       "      <td>IAH</td>\n",
       "      <td>...</td>\n",
       "      <td>964</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2019-03-19</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>151</td>\n",
       "      <td>AA</td>\n",
       "      <td>N868NN</td>\n",
       "      <td>151</td>\n",
       "      <td>14492</td>\n",
       "      <td>RDU</td>\n",
       "      <td>...</td>\n",
       "      <td>130</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      fl_date mkt_unique_carrier branded_code_share mkt_carrier  \\\n",
       "0  2019-11-13                 AA       AA_CODESHARE          AA   \n",
       "1  2019-12-08                 B6                 B6          B6   \n",
       "2  2018-08-01                 UA       UA_CODESHARE          UA   \n",
       "3  2018-02-27                 WN                 WN          WN   \n",
       "4  2019-05-27                 WN                 WN          WN   \n",
       "5  2019-06-12                 AA       AA_CODESHARE          AA   \n",
       "6  2018-06-24                 DL                 DL          DL   \n",
       "7  2019-10-18                 WN                 WN          WN   \n",
       "8  2019-03-30                 AA                 AA          AA   \n",
       "9  2019-03-19                 AA                 AA          AA   \n",
       "\n",
       "   mkt_carrier_fl_num op_unique_carrier tail_num  op_carrier_fl_num  \\\n",
       "0                4850                PT   N657AE               4850   \n",
       "1                 168                B6   N633JB                168   \n",
       "2                5068                OO   N975SW               5068   \n",
       "3                 388                WN   N968WN                388   \n",
       "4                1353                WN   N8575Z               1353   \n",
       "5                3696                MQ   N667GB               3696   \n",
       "6                1180                DL   N909DA               1180   \n",
       "7                2277                WN   N209WN               2277   \n",
       "8                 711                AA   N818AW                711   \n",
       "9                 151                AA   N868NN                151   \n",
       "\n",
       "   origin_airport_id origin  ... distance  carrier_delay weather_delay  \\\n",
       "0              15096    SYR  ...      228            NaN           NaN   \n",
       "1              13796    OAK  ...     2576            NaN           NaN   \n",
       "2              11027    CKB  ...      151            NaN           NaN   \n",
       "3              10821    BWI  ...      430           48.0           0.0   \n",
       "4              12339    IND  ...      977            NaN           NaN   \n",
       "5              13930    ORD  ...      135            0.0           0.0   \n",
       "6              11193    CVG  ...     1069            NaN           NaN   \n",
       "7              13296    MHT  ...      377            NaN           NaN   \n",
       "8              12266    IAH  ...      964            NaN           NaN   \n",
       "9              14492    RDU  ...      130            NaN           NaN   \n",
       "\n",
       "  nas_delay  security_delay  late_aircraft_delay  first_dep_time  \\\n",
       "0       NaN             NaN                  NaN             NaN   \n",
       "1       NaN             NaN                  NaN             NaN   \n",
       "2       NaN             NaN                  NaN             NaN   \n",
       "3       0.0             0.0                  0.0             NaN   \n",
       "4       NaN             NaN                  NaN             NaN   \n",
       "5      36.0             0.0                 18.0             NaN   \n",
       "6       NaN             NaN                  NaN             NaN   \n",
       "7       NaN             NaN                  NaN             NaN   \n",
       "8       NaN             NaN                  NaN             NaN   \n",
       "9       NaN             NaN                  NaN             NaN   \n",
       "\n",
       "   total_add_gtime  longest_add_gtime  no_name  \n",
       "0              NaN                NaN      NaN  \n",
       "1              NaN                NaN      NaN  \n",
       "2              NaN                NaN      NaN  \n",
       "3              NaN                NaN      NaN  \n",
       "4              NaN                NaN      NaN  \n",
       "5              NaN                NaN      NaN  \n",
       "6              NaN                NaN      NaN  \n",
       "7              NaN                NaN      NaN  \n",
       "8              NaN                NaN      NaN  \n",
       "9              NaN                NaN      NaN  \n",
       "\n",
       "[10 rows x 42 columns]"
      ]
     },
     "execution_count": 718,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Loading the trainning dataset\n",
    "\n",
    "train_flights_subset = pd.read_csv(\"/Users/rafaelaqueiroz/Mid-Term-Project-I/train_flights_subset.csv\", sep = \",\")\n",
    "train_flights_subset.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 719,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7000, 42)\n",
      "fl_date                 object\n",
      "mkt_unique_carrier      object\n",
      "branded_code_share      object\n",
      "mkt_carrier             object\n",
      "mkt_carrier_fl_num       int64\n",
      "op_unique_carrier       object\n",
      "tail_num                object\n",
      "op_carrier_fl_num        int64\n",
      "origin_airport_id        int64\n",
      "origin                  object\n",
      "origin_city_name        object\n",
      "dest_airport_id          int64\n",
      "dest                    object\n",
      "dest_city_name          object\n",
      "crs_dep_time             int64\n",
      "dep_time               float64\n",
      "dep_delay              float64\n",
      "taxi_out               float64\n",
      "wheels_off             float64\n",
      "wheels_on              float64\n",
      "taxi_in                float64\n",
      "crs_arr_time             int64\n",
      "arr_time               float64\n",
      "arr_delay              float64\n",
      "cancelled                int64\n",
      "cancellation_code       object\n",
      "diverted                 int64\n",
      "dup                     object\n",
      "crs_elapsed_time         int64\n",
      "actual_elapsed_time    float64\n",
      "air_time               float64\n",
      "flights                  int64\n",
      "distance                 int64\n",
      "carrier_delay          float64\n",
      "weather_delay          float64\n",
      "nas_delay              float64\n",
      "security_delay         float64\n",
      "late_aircraft_delay    float64\n",
      "first_dep_time         float64\n",
      "total_add_gtime        float64\n",
      "longest_add_gtime      float64\n",
      "no_name                float64\n",
      "dtype: object\n"
     ]
    }
   ],
   "source": [
    "# Checking the shape, the columns and dtypes\n",
    "\n",
    "print(train_flights_subset.shape)\n",
    "print(train_flights_subset.dtypes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 720,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fl_date                   0\n",
       "mkt_unique_carrier        0\n",
       "branded_code_share        0\n",
       "mkt_carrier               0\n",
       "mkt_carrier_fl_num        0\n",
       "op_unique_carrier         0\n",
       "tail_num                 25\n",
       "op_carrier_fl_num         0\n",
       "origin_airport_id         0\n",
       "origin                    0\n",
       "origin_city_name          0\n",
       "dest_airport_id           0\n",
       "dest                      0\n",
       "dest_city_name            0\n",
       "crs_dep_time              0\n",
       "dep_time                111\n",
       "dep_delay               111\n",
       "taxi_out                120\n",
       "wheels_off              120\n",
       "wheels_on               122\n",
       "taxi_in                 122\n",
       "crs_arr_time              0\n",
       "arr_time                119\n",
       "arr_delay               132\n",
       "cancelled                 0\n",
       "cancellation_code      6883\n",
       "diverted                  0\n",
       "dup                       0\n",
       "crs_elapsed_time          0\n",
       "actual_elapsed_time     131\n",
       "air_time                134\n",
       "flights                   0\n",
       "distance                  0\n",
       "carrier_delay          5703\n",
       "weather_delay          5703\n",
       "nas_delay              5703\n",
       "security_delay         5703\n",
       "late_aircraft_delay    5703\n",
       "first_dep_time         6941\n",
       "total_add_gtime        6941\n",
       "longest_add_gtime      6941\n",
       "no_name                7000\n",
       "dtype: int64"
      ]
     },
     "execution_count": 720,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking for missing values and NaN\n",
    "\n",
    "train_flights_subset.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 721,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fl_date</th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>branded_code_share</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>...</th>\n",
       "      <th>distance</th>\n",
       "      <th>carrier_delay</th>\n",
       "      <th>weather_delay</th>\n",
       "      <th>nas_delay</th>\n",
       "      <th>security_delay</th>\n",
       "      <th>late_aircraft_delay</th>\n",
       "      <th>first_dep_time</th>\n",
       "      <th>total_add_gtime</th>\n",
       "      <th>longest_add_gtime</th>\n",
       "      <th>no_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2019-11-13</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA_CODESHARE</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>...</td>\n",
       "      <td>228</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2019-12-08</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>...</td>\n",
       "      <td>2576</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>UA</td>\n",
       "      <td>UA_CODESHARE</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>...</td>\n",
       "      <td>151</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-02-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>...</td>\n",
       "      <td>430</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019-05-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>...</td>\n",
       "      <td>977</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6995</th>\n",
       "      <td>2018-03-26</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>850</td>\n",
       "      <td>AA</td>\n",
       "      <td>N197UW</td>\n",
       "      <td>850</td>\n",
       "      <td>13303</td>\n",
       "      <td>MIA</td>\n",
       "      <td>...</td>\n",
       "      <td>1013</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6996</th>\n",
       "      <td>2019-02-05</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>2665</td>\n",
       "      <td>AA</td>\n",
       "      <td>N196AA</td>\n",
       "      <td>2665</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>...</td>\n",
       "      <td>1013</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6997</th>\n",
       "      <td>2018-12-26</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA_CODESHARE</td>\n",
       "      <td>AA</td>\n",
       "      <td>5180</td>\n",
       "      <td>OH</td>\n",
       "      <td>N725PS</td>\n",
       "      <td>5180</td>\n",
       "      <td>11057</td>\n",
       "      <td>CLT</td>\n",
       "      <td>...</td>\n",
       "      <td>564</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6998</th>\n",
       "      <td>2018-11-10</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>576</td>\n",
       "      <td>AA</td>\n",
       "      <td>N188US</td>\n",
       "      <td>576</td>\n",
       "      <td>14893</td>\n",
       "      <td>SMF</td>\n",
       "      <td>...</td>\n",
       "      <td>2244</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6999</th>\n",
       "      <td>2018-05-03</td>\n",
       "      <td>DL</td>\n",
       "      <td>DL_CODESHARE</td>\n",
       "      <td>DL</td>\n",
       "      <td>4669</td>\n",
       "      <td>OO</td>\n",
       "      <td>N772SK</td>\n",
       "      <td>4669</td>\n",
       "      <td>13487</td>\n",
       "      <td>MSP</td>\n",
       "      <td>...</td>\n",
       "      <td>596</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>1304.79661</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>36.508475</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7000 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         fl_date mkt_unique_carrier branded_code_share mkt_carrier  \\\n",
       "0     2019-11-13                 AA       AA_CODESHARE          AA   \n",
       "1     2019-12-08                 B6                 B6          B6   \n",
       "2     2018-08-01                 UA       UA_CODESHARE          UA   \n",
       "3     2018-02-27                 WN                 WN          WN   \n",
       "4     2019-05-27                 WN                 WN          WN   \n",
       "...          ...                ...                ...         ...   \n",
       "6995  2018-03-26                 AA                 AA          AA   \n",
       "6996  2019-02-05                 AA                 AA          AA   \n",
       "6997  2018-12-26                 AA       AA_CODESHARE          AA   \n",
       "6998  2018-11-10                 AA                 AA          AA   \n",
       "6999  2018-05-03                 DL       DL_CODESHARE          DL   \n",
       "\n",
       "      mkt_carrier_fl_num op_unique_carrier tail_num  op_carrier_fl_num  \\\n",
       "0                   4850                PT   N657AE               4850   \n",
       "1                    168                B6   N633JB                168   \n",
       "2                   5068                OO   N975SW               5068   \n",
       "3                    388                WN   N968WN                388   \n",
       "4                   1353                WN   N8575Z               1353   \n",
       "...                  ...               ...      ...                ...   \n",
       "6995                 850                AA   N197UW                850   \n",
       "6996                2665                AA   N196AA               2665   \n",
       "6997                5180                OH   N725PS               5180   \n",
       "6998                 576                AA   N188US                576   \n",
       "6999                4669                OO   N772SK               4669   \n",
       "\n",
       "      origin_airport_id origin  ... distance  carrier_delay weather_delay  \\\n",
       "0                 15096    SYR  ...      228      19.782575      2.678489   \n",
       "1                 13796    OAK  ...     2576      19.782575      2.678489   \n",
       "2                 11027    CKB  ...      151      19.782575      2.678489   \n",
       "3                 10821    BWI  ...      430      48.000000      0.000000   \n",
       "4                 12339    IND  ...      977      19.782575      2.678489   \n",
       "...                 ...    ...  ...      ...            ...           ...   \n",
       "6995              13303    MIA  ...     1013      19.782575      2.678489   \n",
       "6996              14100    PHL  ...     1013      19.782575      2.678489   \n",
       "6997              11057    CLT  ...      564      19.782575      2.678489   \n",
       "6998              14893    SMF  ...     2244      19.782575      2.678489   \n",
       "6999              13487    MSP  ...      596      19.782575      2.678489   \n",
       "\n",
       "      nas_delay  security_delay  late_aircraft_delay  first_dep_time  \\\n",
       "0     17.033153        0.087124            26.690825      1304.79661   \n",
       "1     17.033153        0.087124            26.690825      1304.79661   \n",
       "2     17.033153        0.087124            26.690825      1304.79661   \n",
       "3      0.000000        0.000000             0.000000      1304.79661   \n",
       "4     17.033153        0.087124            26.690825      1304.79661   \n",
       "...         ...             ...                  ...             ...   \n",
       "6995  17.033153        0.087124            26.690825      1304.79661   \n",
       "6996  17.033153        0.087124            26.690825      1304.79661   \n",
       "6997  17.033153        0.087124            26.690825      1304.79661   \n",
       "6998  17.033153        0.087124            26.690825      1304.79661   \n",
       "6999  17.033153        0.087124            26.690825      1304.79661   \n",
       "\n",
       "      total_add_gtime  longest_add_gtime  no_name  \n",
       "0           36.508475          36.508475      NaN  \n",
       "1           36.508475          36.508475      NaN  \n",
       "2           36.508475          36.508475      NaN  \n",
       "3           36.508475          36.508475      NaN  \n",
       "4           36.508475          36.508475      NaN  \n",
       "...               ...                ...      ...  \n",
       "6995        36.508475          36.508475      NaN  \n",
       "6996        36.508475          36.508475      NaN  \n",
       "6997        36.508475          36.508475      NaN  \n",
       "6998        36.508475          36.508475      NaN  \n",
       "6999        36.508475          36.508475      NaN  \n",
       "\n",
       "[7000 rows x 42 columns]"
      ]
     },
     "execution_count": 721,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Handling missing values with the mean\n",
    "\n",
    "train_flights_subset.fillna(train_flights_subset.mean(), inplace=True)\n",
    "train_flights_subset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 722,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fl_date</th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>branded_code_share</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>...</th>\n",
       "      <th>distance</th>\n",
       "      <th>carrier_delay</th>\n",
       "      <th>weather_delay</th>\n",
       "      <th>nas_delay</th>\n",
       "      <th>security_delay</th>\n",
       "      <th>late_aircraft_delay</th>\n",
       "      <th>first_dep_time</th>\n",
       "      <th>total_add_gtime</th>\n",
       "      <th>longest_add_gtime</th>\n",
       "      <th>no_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6995</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6996</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6997</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6998</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6999</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7000 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      fl_date  mkt_unique_carrier  branded_code_share  mkt_carrier  \\\n",
       "0       False               False               False        False   \n",
       "1       False               False               False        False   \n",
       "2       False               False               False        False   \n",
       "3       False               False               False        False   \n",
       "4       False               False               False        False   \n",
       "...       ...                 ...                 ...          ...   \n",
       "6995    False               False               False        False   \n",
       "6996    False               False               False        False   \n",
       "6997    False               False               False        False   \n",
       "6998    False               False               False        False   \n",
       "6999    False               False               False        False   \n",
       "\n",
       "      mkt_carrier_fl_num  op_unique_carrier  tail_num  op_carrier_fl_num  \\\n",
       "0                  False              False     False              False   \n",
       "1                  False              False     False              False   \n",
       "2                  False              False     False              False   \n",
       "3                  False              False     False              False   \n",
       "4                  False              False     False              False   \n",
       "...                  ...                ...       ...                ...   \n",
       "6995               False              False     False              False   \n",
       "6996               False              False     False              False   \n",
       "6997               False              False     False              False   \n",
       "6998               False              False     False              False   \n",
       "6999               False              False     False              False   \n",
       "\n",
       "      origin_airport_id  origin  ...  distance  carrier_delay  weather_delay  \\\n",
       "0                 False   False  ...     False          False          False   \n",
       "1                 False   False  ...     False          False          False   \n",
       "2                 False   False  ...     False          False          False   \n",
       "3                 False   False  ...     False          False          False   \n",
       "4                 False   False  ...     False          False          False   \n",
       "...                 ...     ...  ...       ...            ...            ...   \n",
       "6995              False   False  ...     False          False          False   \n",
       "6996              False   False  ...     False          False          False   \n",
       "6997              False   False  ...     False          False          False   \n",
       "6998              False   False  ...     False          False          False   \n",
       "6999              False   False  ...     False          False          False   \n",
       "\n",
       "      nas_delay  security_delay  late_aircraft_delay  first_dep_time  \\\n",
       "0         False           False                False           False   \n",
       "1         False           False                False           False   \n",
       "2         False           False                False           False   \n",
       "3         False           False                False           False   \n",
       "4         False           False                False           False   \n",
       "...         ...             ...                  ...             ...   \n",
       "6995      False           False                False           False   \n",
       "6996      False           False                False           False   \n",
       "6997      False           False                False           False   \n",
       "6998      False           False                False           False   \n",
       "6999      False           False                False           False   \n",
       "\n",
       "      total_add_gtime  longest_add_gtime  no_name  \n",
       "0               False              False     True  \n",
       "1               False              False     True  \n",
       "2               False              False     True  \n",
       "3               False              False     True  \n",
       "4               False              False     True  \n",
       "...               ...                ...      ...  \n",
       "6995            False              False     True  \n",
       "6996            False              False     True  \n",
       "6997            False              False     True  \n",
       "6998            False              False     True  \n",
       "6999            False              False     True  \n",
       "\n",
       "[7000 rows x 42 columns]"
      ]
     },
     "execution_count": 722,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking if there are still missing values in the df\n",
    "\n",
    "train_flights_subset.isnull()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 723,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fl_date</th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>...</th>\n",
       "      <th>crs_elapsed_time</th>\n",
       "      <th>actual_elapsed_time</th>\n",
       "      <th>air_time</th>\n",
       "      <th>flights</th>\n",
       "      <th>distance</th>\n",
       "      <th>carrier_delay</th>\n",
       "      <th>weather_delay</th>\n",
       "      <th>nas_delay</th>\n",
       "      <th>security_delay</th>\n",
       "      <th>late_aircraft_delay</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2019-11-13</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>...</td>\n",
       "      <td>84</td>\n",
       "      <td>62.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>1</td>\n",
       "      <td>228</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2019-12-08</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>...</td>\n",
       "      <td>319</td>\n",
       "      <td>330.0</td>\n",
       "      <td>302.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2576</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>...</td>\n",
       "      <td>64</td>\n",
       "      <td>47.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>1</td>\n",
       "      <td>151</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-02-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>...</td>\n",
       "      <td>105</td>\n",
       "      <td>92.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>1</td>\n",
       "      <td>430</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019-05-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>...</td>\n",
       "      <td>165</td>\n",
       "      <td>173.0</td>\n",
       "      <td>157.0</td>\n",
       "      <td>1</td>\n",
       "      <td>977</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6995</th>\n",
       "      <td>2018-03-26</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>850</td>\n",
       "      <td>AA</td>\n",
       "      <td>N197UW</td>\n",
       "      <td>850</td>\n",
       "      <td>13303</td>\n",
       "      <td>MIA</td>\n",
       "      <td>Miami, FL</td>\n",
       "      <td>...</td>\n",
       "      <td>171</td>\n",
       "      <td>169.0</td>\n",
       "      <td>147.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1013</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6996</th>\n",
       "      <td>2019-02-05</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>2665</td>\n",
       "      <td>AA</td>\n",
       "      <td>N196AA</td>\n",
       "      <td>2665</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>Philadelphia, PA</td>\n",
       "      <td>...</td>\n",
       "      <td>186</td>\n",
       "      <td>157.0</td>\n",
       "      <td>137.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1013</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6997</th>\n",
       "      <td>2018-12-26</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>5180</td>\n",
       "      <td>OH</td>\n",
       "      <td>N725PS</td>\n",
       "      <td>5180</td>\n",
       "      <td>11057</td>\n",
       "      <td>CLT</td>\n",
       "      <td>Charlotte, NC</td>\n",
       "      <td>...</td>\n",
       "      <td>115</td>\n",
       "      <td>102.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>1</td>\n",
       "      <td>564</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6998</th>\n",
       "      <td>2018-11-10</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>576</td>\n",
       "      <td>AA</td>\n",
       "      <td>N188US</td>\n",
       "      <td>576</td>\n",
       "      <td>14893</td>\n",
       "      <td>SMF</td>\n",
       "      <td>Sacramento, CA</td>\n",
       "      <td>...</td>\n",
       "      <td>273</td>\n",
       "      <td>256.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2244</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6999</th>\n",
       "      <td>2018-05-03</td>\n",
       "      <td>DL</td>\n",
       "      <td>DL</td>\n",
       "      <td>4669</td>\n",
       "      <td>OO</td>\n",
       "      <td>N772SK</td>\n",
       "      <td>4669</td>\n",
       "      <td>13487</td>\n",
       "      <td>MSP</td>\n",
       "      <td>Minneapolis, MN</td>\n",
       "      <td>...</td>\n",
       "      <td>118</td>\n",
       "      <td>109.0</td>\n",
       "      <td>83.0</td>\n",
       "      <td>1</td>\n",
       "      <td>596</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7000 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         fl_date mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num  \\\n",
       "0     2019-11-13                 AA          AA                4850   \n",
       "1     2019-12-08                 B6          B6                 168   \n",
       "2     2018-08-01                 UA          UA                5068   \n",
       "3     2018-02-27                 WN          WN                 388   \n",
       "4     2019-05-27                 WN          WN                1353   \n",
       "...          ...                ...         ...                 ...   \n",
       "6995  2018-03-26                 AA          AA                 850   \n",
       "6996  2019-02-05                 AA          AA                2665   \n",
       "6997  2018-12-26                 AA          AA                5180   \n",
       "6998  2018-11-10                 AA          AA                 576   \n",
       "6999  2018-05-03                 DL          DL                4669   \n",
       "\n",
       "     op_unique_carrier tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0                   PT   N657AE               4850              15096    SYR   \n",
       "1                   B6   N633JB                168              13796    OAK   \n",
       "2                   OO   N975SW               5068              11027    CKB   \n",
       "3                   WN   N968WN                388              10821    BWI   \n",
       "4                   WN   N8575Z               1353              12339    IND   \n",
       "...                ...      ...                ...                ...    ...   \n",
       "6995                AA   N197UW                850              13303    MIA   \n",
       "6996                AA   N196AA               2665              14100    PHL   \n",
       "6997                OH   N725PS               5180              11057    CLT   \n",
       "6998                AA   N188US                576              14893    SMF   \n",
       "6999                OO   N772SK               4669              13487    MSP   \n",
       "\n",
       "             origin_city_name  ...  crs_elapsed_time actual_elapsed_time  \\\n",
       "0                Syracuse, NY  ...                84                62.0   \n",
       "1                 Oakland, CA  ...               319               330.0   \n",
       "2     Clarksburg/Fairmont, WV  ...                64                47.0   \n",
       "3               Baltimore, MD  ...               105                92.0   \n",
       "4            Indianapolis, IN  ...               165               173.0   \n",
       "...                       ...  ...               ...                 ...   \n",
       "6995                Miami, FL  ...               171               169.0   \n",
       "6996         Philadelphia, PA  ...               186               157.0   \n",
       "6997            Charlotte, NC  ...               115               102.0   \n",
       "6998           Sacramento, CA  ...               273               256.0   \n",
       "6999          Minneapolis, MN  ...               118               109.0   \n",
       "\n",
       "     air_time  flights  distance  carrier_delay  weather_delay  nas_delay  \\\n",
       "0        48.0        1       228      19.782575       2.678489  17.033153   \n",
       "1       302.0        1      2576      19.782575       2.678489  17.033153   \n",
       "2        30.0        1       151      19.782575       2.678489  17.033153   \n",
       "3        74.0        1       430      48.000000       0.000000   0.000000   \n",
       "4       157.0        1       977      19.782575       2.678489  17.033153   \n",
       "...       ...      ...       ...            ...            ...        ...   \n",
       "6995    147.0        1      1013      19.782575       2.678489  17.033153   \n",
       "6996    137.0        1      1013      19.782575       2.678489  17.033153   \n",
       "6997     91.0        1       564      19.782575       2.678489  17.033153   \n",
       "6998    238.0        1      2244      19.782575       2.678489  17.033153   \n",
       "6999     83.0        1       596      19.782575       2.678489  17.033153   \n",
       "\n",
       "     security_delay  late_aircraft_delay  \n",
       "0          0.087124            26.690825  \n",
       "1          0.087124            26.690825  \n",
       "2          0.087124            26.690825  \n",
       "3          0.000000             0.000000  \n",
       "4          0.087124            26.690825  \n",
       "...             ...                  ...  \n",
       "6995       0.087124            26.690825  \n",
       "6996       0.087124            26.690825  \n",
       "6997       0.087124            26.690825  \n",
       "6998       0.087124            26.690825  \n",
       "6999       0.087124            26.690825  \n",
       "\n",
       "[7000 rows x 31 columns]"
      ]
     },
     "execution_count": 723,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dropping some columns \n",
    "\n",
    "train_flights_subset = train_flights_subset.drop(['branded_code_share', 'dep_time', 'dep_delay', 'taxi_in', 'taxi_out', 'wheels_off', 'wheels_on', 'no_name', 'longest_add_gtime', 'first_dep_time', 'total_add_gtime'], axis=1)\n",
    "train_flights_subset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 724,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fl_date</th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>...</th>\n",
       "      <th>distance</th>\n",
       "      <th>carrier_delay</th>\n",
       "      <th>weather_delay</th>\n",
       "      <th>nas_delay</th>\n",
       "      <th>security_delay</th>\n",
       "      <th>late_aircraft_delay</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "      <th>day_of_month</th>\n",
       "      <th>day_of_week</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2019-11-13</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>...</td>\n",
       "      <td>228</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2019</td>\n",
       "      <td>11</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2019-12-08</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>...</td>\n",
       "      <td>2576</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2019</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>...</td>\n",
       "      <td>151</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2018</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-02-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>...</td>\n",
       "      <td>430</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2018</td>\n",
       "      <td>2</td>\n",
       "      <td>27</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019-05-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>...</td>\n",
       "      <td>977</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2019</td>\n",
       "      <td>5</td>\n",
       "      <td>27</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     fl_date mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num  \\\n",
       "0 2019-11-13                 AA          AA                4850   \n",
       "1 2019-12-08                 B6          B6                 168   \n",
       "2 2018-08-01                 UA          UA                5068   \n",
       "3 2018-02-27                 WN          WN                 388   \n",
       "4 2019-05-27                 WN          WN                1353   \n",
       "\n",
       "  op_unique_carrier tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0                PT   N657AE               4850              15096    SYR   \n",
       "1                B6   N633JB                168              13796    OAK   \n",
       "2                OO   N975SW               5068              11027    CKB   \n",
       "3                WN   N968WN                388              10821    BWI   \n",
       "4                WN   N8575Z               1353              12339    IND   \n",
       "\n",
       "          origin_city_name  ...  distance carrier_delay weather_delay  \\\n",
       "0             Syracuse, NY  ...       228     19.782575      2.678489   \n",
       "1              Oakland, CA  ...      2576     19.782575      2.678489   \n",
       "2  Clarksburg/Fairmont, WV  ...       151     19.782575      2.678489   \n",
       "3            Baltimore, MD  ...       430     48.000000      0.000000   \n",
       "4         Indianapolis, IN  ...       977     19.782575      2.678489   \n",
       "\n",
       "   nas_delay  security_delay  late_aircraft_delay  year  month day_of_month  \\\n",
       "0  17.033153        0.087124            26.690825  2019     11           13   \n",
       "1  17.033153        0.087124            26.690825  2019     12            8   \n",
       "2  17.033153        0.087124            26.690825  2018      8            1   \n",
       "3   0.000000        0.000000             0.000000  2018      2           27   \n",
       "4  17.033153        0.087124            26.690825  2019      5           27   \n",
       "\n",
       "   day_of_week  \n",
       "0            2  \n",
       "1            6  \n",
       "2            2  \n",
       "3            1  \n",
       "4            0  \n",
       "\n",
       "[5 rows x 35 columns]"
      ]
     },
     "execution_count": 724,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Starting first by extracting the date time to explore this variable\n",
    "\n",
    "train_flights_subset['fl_date'] = pd.to_datetime(train_flights_subset['fl_date'], errors='coerce')\n",
    "train_flights_subset['year'] = train_flights_subset['fl_date'].dt.year\n",
    "train_flights_subset['month'] = train_flights_subset['fl_date'].dt.month\n",
    "train_flights_subset['day_of_month'] = train_flights_subset['fl_date'].dt.day\n",
    "train_flights_subset['day_of_week'] = train_flights_subset['fl_date'].dt.dayofweek\n",
    "train_flights_subset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 725,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fl_date</th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>...</th>\n",
       "      <th>weather_delay</th>\n",
       "      <th>nas_delay</th>\n",
       "      <th>security_delay</th>\n",
       "      <th>late_aircraft_delay</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "      <th>day_of_month</th>\n",
       "      <th>day_of_week</th>\n",
       "      <th>quarter</th>\n",
       "      <th>semester</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2019-11-13</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>...</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2019</td>\n",
       "      <td>11</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2019-12-08</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>...</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2019</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>...</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2018</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>Q3</td>\n",
       "      <td>S3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-02-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2018</td>\n",
       "      <td>2</td>\n",
       "      <td>27</td>\n",
       "      <td>1</td>\n",
       "      <td>Q1</td>\n",
       "      <td>S1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019-05-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>...</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>2019</td>\n",
       "      <td>5</td>\n",
       "      <td>27</td>\n",
       "      <td>0</td>\n",
       "      <td>Q2</td>\n",
       "      <td>S2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 37 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     fl_date mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num  \\\n",
       "0 2019-11-13                 AA          AA                4850   \n",
       "1 2019-12-08                 B6          B6                 168   \n",
       "2 2018-08-01                 UA          UA                5068   \n",
       "3 2018-02-27                 WN          WN                 388   \n",
       "4 2019-05-27                 WN          WN                1353   \n",
       "\n",
       "  op_unique_carrier tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0                PT   N657AE               4850              15096    SYR   \n",
       "1                B6   N633JB                168              13796    OAK   \n",
       "2                OO   N975SW               5068              11027    CKB   \n",
       "3                WN   N968WN                388              10821    BWI   \n",
       "4                WN   N8575Z               1353              12339    IND   \n",
       "\n",
       "          origin_city_name  ...  weather_delay  nas_delay security_delay  \\\n",
       "0             Syracuse, NY  ...       2.678489  17.033153       0.087124   \n",
       "1              Oakland, CA  ...       2.678489  17.033153       0.087124   \n",
       "2  Clarksburg/Fairmont, WV  ...       2.678489  17.033153       0.087124   \n",
       "3            Baltimore, MD  ...       0.000000   0.000000       0.000000   \n",
       "4         Indianapolis, IN  ...       2.678489  17.033153       0.087124   \n",
       "\n",
       "   late_aircraft_delay  year  month  day_of_month  day_of_week quarter  \\\n",
       "0            26.690825  2019     11            13            2      Q4   \n",
       "1            26.690825  2019     12             8            6      Q4   \n",
       "2            26.690825  2018      8             1            2      Q3   \n",
       "3             0.000000  2018      2            27            1      Q1   \n",
       "4            26.690825  2019      5            27            0      Q2   \n",
       "\n",
       "   semester  \n",
       "0        S4  \n",
       "1        S4  \n",
       "2        S3  \n",
       "3        S1  \n",
       "4        S2  \n",
       "\n",
       "[5 rows x 37 columns]"
      ]
     },
     "execution_count": 725,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Besides extracting the information about the year, month, date_of_month and day_of_week above, the seasonality or semester can also give us information to help us with delays predictions\n",
    "# Adding quarter and semester columns with periodindex and strftime functions\n",
    "# We are using the calendar year starting in January\n",
    "\n",
    "train_flights_subset['quarter'] = pd.PeriodIndex(train_flights_subset['fl_date'], freq='Q').strftime('Q%q')\n",
    "train_flights_subset['semester'] = pd.PeriodIndex(train_flights_subset['fl_date'], freq='Q').strftime('S%q')\n",
    "train_flights_subset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 726,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Finally, let's add a time of day column into our df as traffic can vary from day to night\n",
    "# Firts we are creating a function to get the time of the day\n",
    "# Then we are using the crs_dep_time as this is the scheduled departure time of the flight so it is more accurate and after the crs_arr_time as this is the sceduled arrival time of the plane\n",
    "\n",
    "def get_time_of_day(hour):\n",
    "    if hour < 6:\n",
    "        return \"night\"\n",
    "    elif hour < 12:\n",
    "        return \"morning\"\n",
    "    elif hour < 17:\n",
    "        return \"afternoon\"\n",
    "    else:\n",
    "        return \"evening\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 727,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fl_date</th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>...</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "      <th>day_of_month</th>\n",
       "      <th>day_of_week</th>\n",
       "      <th>quarter</th>\n",
       "      <th>semester</th>\n",
       "      <th>dep_hour_of_day</th>\n",
       "      <th>dep_time_of_day</th>\n",
       "      <th>arr_hour_of_day</th>\n",
       "      <th>arr_time_of_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2019-11-13</td>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>...</td>\n",
       "      <td>2019</td>\n",
       "      <td>11</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>14</td>\n",
       "      <td>afternoon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2019-12-08</td>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>...</td>\n",
       "      <td>2019</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>23</td>\n",
       "      <td>evening</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>Q3</td>\n",
       "      <td>S3</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-02-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>2</td>\n",
       "      <td>27</td>\n",
       "      <td>1</td>\n",
       "      <td>Q1</td>\n",
       "      <td>S1</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019-05-27</td>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>...</td>\n",
       "      <td>2019</td>\n",
       "      <td>5</td>\n",
       "      <td>27</td>\n",
       "      <td>0</td>\n",
       "      <td>Q2</td>\n",
       "      <td>S2</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 41 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     fl_date mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num  \\\n",
       "0 2019-11-13                 AA          AA                4850   \n",
       "1 2019-12-08                 B6          B6                 168   \n",
       "2 2018-08-01                 UA          UA                5068   \n",
       "3 2018-02-27                 WN          WN                 388   \n",
       "4 2019-05-27                 WN          WN                1353   \n",
       "\n",
       "  op_unique_carrier tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0                PT   N657AE               4850              15096    SYR   \n",
       "1                B6   N633JB                168              13796    OAK   \n",
       "2                OO   N975SW               5068              11027    CKB   \n",
       "3                WN   N968WN                388              10821    BWI   \n",
       "4                WN   N8575Z               1353              12339    IND   \n",
       "\n",
       "          origin_city_name  ...  year month day_of_month  day_of_week  \\\n",
       "0             Syracuse, NY  ...  2019    11           13            2   \n",
       "1              Oakland, CA  ...  2019    12            8            6   \n",
       "2  Clarksburg/Fairmont, WV  ...  2018     8            1            2   \n",
       "3            Baltimore, MD  ...  2018     2           27            1   \n",
       "4         Indianapolis, IN  ...  2019     5           27            0   \n",
       "\n",
       "   quarter  semester  dep_hour_of_day  dep_time_of_day arr_hour_of_day  \\\n",
       "0       Q4        S4               13        afternoon              14   \n",
       "1       Q4        S4               23          evening               8   \n",
       "2       Q3        S3                6          morning               7   \n",
       "3       Q1        S1                8          morning              10   \n",
       "4       Q2        S2                6          morning               7   \n",
       "\n",
       "   arr_time_of_day  \n",
       "0        afternoon  \n",
       "1          morning  \n",
       "2          morning  \n",
       "3          morning  \n",
       "4          morning  \n",
       "\n",
       "[5 rows x 41 columns]"
      ]
     },
     "execution_count": 727,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Getting the official departure time\n",
    "\n",
    "train_flights_subset['dep_hour_of_day'] = train_flights_subset['crs_dep_time'] // 100\n",
    "train_flights_subset['dep_time_of_day'] = train_flights_subset['dep_hour_of_day'].apply(get_time_of_day)\n",
    "\n",
    "# Getting the official arrival time\n",
    "\n",
    "train_flights_subset['arr_hour_of_day'] = train_flights_subset['crs_arr_time'] // 100\n",
    "train_flights_subset['arr_time_of_day'] = train_flights_subset['arr_hour_of_day'].apply(get_time_of_day)\n",
    "train_flights_subset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 728,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7000, 41)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "fl_date                   0\n",
       "mkt_unique_carrier        0\n",
       "mkt_carrier               0\n",
       "mkt_carrier_fl_num        0\n",
       "op_unique_carrier         0\n",
       "tail_num                 25\n",
       "op_carrier_fl_num         0\n",
       "origin_airport_id         0\n",
       "origin                    0\n",
       "origin_city_name          0\n",
       "dest_airport_id           0\n",
       "dest                      0\n",
       "dest_city_name            0\n",
       "crs_dep_time              0\n",
       "crs_arr_time              0\n",
       "arr_time                  0\n",
       "arr_delay                 0\n",
       "cancelled                 0\n",
       "cancellation_code      6883\n",
       "diverted                  0\n",
       "dup                       0\n",
       "crs_elapsed_time          0\n",
       "actual_elapsed_time       0\n",
       "air_time                  0\n",
       "flights                   0\n",
       "distance                  0\n",
       "carrier_delay             0\n",
       "weather_delay             0\n",
       "nas_delay                 0\n",
       "security_delay            0\n",
       "late_aircraft_delay       0\n",
       "year                      0\n",
       "month                     0\n",
       "day_of_month              0\n",
       "day_of_week               0\n",
       "quarter                   0\n",
       "semester                  0\n",
       "dep_hour_of_day           0\n",
       "dep_time_of_day           0\n",
       "arr_hour_of_day           0\n",
       "arr_time_of_day           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 728,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking the columns list and if there is still any missing value\n",
    "\n",
    "print(train_flights_subset.shape)\n",
    "train_flights_subset.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 729,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>...</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "      <th>day_of_month</th>\n",
       "      <th>day_of_week</th>\n",
       "      <th>quarter</th>\n",
       "      <th>semester</th>\n",
       "      <th>dep_hour_of_day</th>\n",
       "      <th>dep_time_of_day</th>\n",
       "      <th>arr_hour_of_day</th>\n",
       "      <th>arr_time_of_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>2019</td>\n",
       "      <td>11</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>14</td>\n",
       "      <td>afternoon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>12478</td>\n",
       "      <td>...</td>\n",
       "      <td>2019</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>23</td>\n",
       "      <td>evening</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>12264</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>Q3</td>\n",
       "      <td>S3</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>2</td>\n",
       "      <td>27</td>\n",
       "      <td>1</td>\n",
       "      <td>Q1</td>\n",
       "      <td>S1</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>11292</td>\n",
       "      <td>...</td>\n",
       "      <td>2019</td>\n",
       "      <td>5</td>\n",
       "      <td>27</td>\n",
       "      <td>0</td>\n",
       "      <td>Q2</td>\n",
       "      <td>S2</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num op_unique_carrier  \\\n",
       "0                 AA          AA                4850                PT   \n",
       "1                 B6          B6                 168                B6   \n",
       "2                 UA          UA                5068                OO   \n",
       "3                 WN          WN                 388                WN   \n",
       "4                 WN          WN                1353                WN   \n",
       "\n",
       "  tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0   N657AE               4850              15096    SYR   \n",
       "1   N633JB                168              13796    OAK   \n",
       "2   N975SW               5068              11027    CKB   \n",
       "3   N968WN                388              10821    BWI   \n",
       "4   N8575Z               1353              12339    IND   \n",
       "\n",
       "          origin_city_name  dest_airport_id  ...  year month  day_of_month  \\\n",
       "0             Syracuse, NY            14100  ...  2019    11            13   \n",
       "1              Oakland, CA            12478  ...  2019    12             8   \n",
       "2  Clarksburg/Fairmont, WV            12264  ...  2018     8             1   \n",
       "3            Baltimore, MD            11193  ...  2018     2            27   \n",
       "4         Indianapolis, IN            11292  ...  2019     5            27   \n",
       "\n",
       "   day_of_week quarter  semester  dep_hour_of_day  dep_time_of_day  \\\n",
       "0            2      Q4        S4               13        afternoon   \n",
       "1            6      Q4        S4               23          evening   \n",
       "2            2      Q3        S3                6          morning   \n",
       "3            1      Q1        S1                8          morning   \n",
       "4            0      Q2        S2                6          morning   \n",
       "\n",
       "   arr_hour_of_day  arr_time_of_day  \n",
       "0               14        afternoon  \n",
       "1                8          morning  \n",
       "2                7          morning  \n",
       "3               10          morning  \n",
       "4                7          morning  \n",
       "\n",
       "[5 rows x 35 columns]"
      ]
     },
     "execution_count": 729,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dropping more columns to move forward \n",
    "\n",
    "train_flights_subset = train_flights_subset.drop(['fl_date', 'cancelled', 'cancellation_code', 'arr_time', 'crs_dep_time', 'crs_arr_time'], axis = 1)\n",
    "\n",
    "# Dropping the 25 missing number rows of the airplane tail_num\n",
    "\n",
    "train_flights_subset.dropna(subset=['tail_num'], inplace=True)\n",
    "train_flights_subset.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we could extract the information per year, month, week, day, semester and time of the day, we are going to check the counts of flights delayed by each piece of information to understand if there is a pattern and if this information make sense before modelling."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 730,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dep_time_of_day\n",
       "afternoon    825\n",
       "evening      865\n",
       "morning      794\n",
       "night         59\n",
       "Name: arr_delay, dtype: int64"
      ]
     },
     "execution_count": 730,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Using the groupby() method to group the data by each feature (day, week, month, quarter, time of the day) \n",
    "# Counting the number of delayed flights in each group using the sum() method\n",
    "# Our target variable 'arr_delay' is going to be used \n",
    "# The lambda function (x > 0).sum() counts the number of delayed flights in each group by checking if the arrival delay x is greater than zero and return the sum of it\n",
    "\n",
    "# Grouping by time of the day\n",
    "\n",
    "train_flights_subset_grouped_by_time_counts = train_flights_subset.groupby('dep_time_of_day')['arr_delay'].apply(lambda x: (x > 0).sum())\n",
    "train_flights_subset_grouped_by_time_counts.head(4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 731,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the results of the time of the day\n",
    "\n",
    "# Creating a bar plot of the count of delayed flights by time of day\n",
    "sns.barplot(x=train_flights_subset_grouped_by_time_counts.index, y=train_flights_subset_grouped_by_time_counts.values, palette='Blues')\n",
    "\n",
    "# Setting the x-axis label and tick labels\n",
    "plt.xlabel(None)\n",
    "plt.xticks(rotation=45, ha='right', fontsize=10)\n",
    "\n",
    "# Setting the y-axis label and tick labels\n",
    "plt.ylabel('Count of Delayed Flights', fontsize=10)\n",
    "plt.yticks(fontsize=10)\n",
    "# plt.yticklabels(train_flights_subset_grouped_by_time_counts.values, fontsize=10)\n",
    "\n",
    "# Showing the plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the analysis of the time of the day, we can notice that during the afternoon and evening the airports deal with the highest amount of arrival flight delays."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 732,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "day_of_month\n",
       "1      79\n",
       "2      83\n",
       "3      85\n",
       "4      63\n",
       "5      82\n",
       "6      84\n",
       "7      84\n",
       "8      85\n",
       "9     102\n",
       "10     85\n",
       "11     89\n",
       "12     98\n",
       "13     91\n",
       "14     78\n",
       "15     93\n",
       "16     90\n",
       "17     73\n",
       "18     83\n",
       "19     87\n",
       "20     93\n",
       "21    113\n",
       "22     89\n",
       "23     75\n",
       "24     82\n",
       "25     64\n",
       "26     81\n",
       "27     74\n",
       "28     65\n",
       "29     72\n",
       "30     69\n",
       "31     52\n",
       "Name: arr_delay, dtype: int64"
      ]
     },
     "execution_count": 732,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping by day of month\n",
    "\n",
    "train_flights_subset_grouped_by_day_counts = train_flights_subset.groupby('day_of_month')['arr_delay'].apply(lambda x: (x > 0).sum())\n",
    "train_flights_subset_grouped_by_day_counts.head(31)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 733,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the results of the day of the month\n",
    "\n",
    "sns.barplot(x=train_flights_subset_grouped_by_day_counts.index, y=train_flights_subset_grouped_by_day_counts.values, palette='Blues')\n",
    "\n",
    "plt.xlabel('Days of the Month', fontsize=10)\n",
    "plt.xticks(rotation=45, ha='right', fontsize=10)\n",
    "\n",
    "plt.ylabel('Count of Delayed Flights', fontsize=10)\n",
    "plt.yticks(fontsize=10)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the results described below and graph, we can see that the top 5 days of the month with the highest amount of flight arrival delays are in descending order: 21, 9, 12, 15, 20 and 13."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 734,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "day_of_week\n",
       "0    408\n",
       "1    363\n",
       "2    340\n",
       "3    384\n",
       "4    399\n",
       "5    279\n",
       "6    370\n",
       "Name: arr_delay, dtype: int64"
      ]
     },
     "execution_count": 734,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Groupping by day of the week\n",
    "\n",
    "train_flights_subset_grouped_by_week_counts = train_flights_subset.groupby('day_of_week')['arr_delay'].apply(lambda x: (x > 0).sum())\n",
    "train_flights_subset_grouped_by_week_counts.head(7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the results, we can notice that the week days with the highest count of arrival flight delays are (in descending order): Monday, Friday and Sunday."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 735,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "month\n",
       "1     203\n",
       "2     211\n",
       "3     204\n",
       "4     203\n",
       "5     219\n",
       "6     257\n",
       "7     253\n",
       "8     228\n",
       "9     165\n",
       "10    222\n",
       "11    199\n",
       "12    179\n",
       "Name: arr_delay, dtype: int64"
      ]
     },
     "execution_count": 735,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping by month\n",
    "\n",
    "train_flights_subset_grouped_by_month_counts = train_flights_subset.groupby('month')['arr_delay'].apply(lambda x: (x > 0).sum())\n",
    "train_flights_subset_grouped_by_month_counts.head(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 736,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/zt/gl7xtkgj7238nsqq03s_q7840000gn/T/ipykernel_2344/963365703.py:8: UserWarning: FixedFormatter should only be used together with FixedLocator\n",
      "  ax.set_xticklabels(['January', 'February', 'March', 'April', 'May', 'June', 'July', 'August', 'September', 'October', 'November', 'December'], fontsize=9)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the results\n",
    "\n",
    "fig = plt.figure(figsize=(10, 6))\n",
    "ax = fig.add_subplot(1, 1, 1)\n",
    "\n",
    "ax.bar(train_flights_subset_grouped_by_month_counts.index, train_flights_subset_grouped_by_month_counts.values, color=(0.3, 0.6, 0.9))\n",
    "\n",
    "ax.set_xticklabels(['January', 'February', 'March', 'April', 'May', 'June', 'July', 'August', 'September', 'October', 'November', 'December'], fontsize=9)\n",
    "# ax.set_xlabel('Months', fontsize=12)\n",
    "ax.set_ylabel('Count of Delayed Flights', fontsize=10)\n",
    "\n",
    "plt.xticks(train_flights_subset_grouped_by_month_counts.index)\n",
    "plt.yticks(fontsize=10)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From these results, it is noted that June, July, August and October have the highest count of arrival flight delays. In addition, June and July corresponde to summer season and to many family vacations, therefore these factors combined with others might explain the higher count of delays in those 2 first months."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 737,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "quarter\n",
       "Q1    618\n",
       "Q2    679\n",
       "Q3    646\n",
       "Q4    600\n",
       "Name: arr_delay, dtype: int64"
      ]
     },
     "execution_count": 737,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping by quarter\n",
    "\n",
    "train_flights_subset_grouped_by_quarter_counts = train_flights_subset.groupby('quarter')['arr_delay'].apply(lambda x: (x > 0).sum())\n",
    "train_flights_subset_grouped_by_quarter_counts.head(4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 738,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "semester\n",
       "S1    618\n",
       "S2    679\n",
       "S3    646\n",
       "S4    600\n",
       "Name: arr_delay, dtype: int64"
      ]
     },
     "execution_count": 738,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping by semester\n",
    "\n",
    "train_flights_subset_grouped_by_quarter_counts = train_flights_subset.groupby('semester')['arr_delay'].apply(lambda x: (x > 0).sum())\n",
    "train_flights_subset_grouped_by_quarter_counts.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The results show a similar count of flight delays by quarter, however the Q2 and Q3 represent a little higher amount of flight delays. This information matches with the months that represent a higher count of flight delays presented before (e.g. June, July and August)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 739,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "year\n",
       "2018    1229\n",
       "2019    1314\n",
       "Name: arr_delay, dtype: int64"
      ]
     },
     "execution_count": 739,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping by year\n",
    "\n",
    "train_flights_subset_grouped_by_year_counts = train_flights_subset.groupby('year')['arr_delay'].apply(lambda x: (x > 0).sum())\n",
    "train_flights_subset_grouped_by_year_counts.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Both 2 years represent a similar count of flight arrival delays being the year of 2019 with a higher amount.\n",
    "Obs.: As we are working with a random trainning sample, the results correspond to the years presented on it, in this case, 2018 and 2019."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 740,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>...</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "      <th>day_of_month</th>\n",
       "      <th>day_of_week</th>\n",
       "      <th>quarter</th>\n",
       "      <th>semester</th>\n",
       "      <th>dep_hour_of_day</th>\n",
       "      <th>dep_time_of_day</th>\n",
       "      <th>arr_hour_of_day</th>\n",
       "      <th>arr_time_of_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6995</th>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>850</td>\n",
       "      <td>AA</td>\n",
       "      <td>N197UW</td>\n",
       "      <td>850</td>\n",
       "      <td>13303</td>\n",
       "      <td>MIA</td>\n",
       "      <td>Miami, FL</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>3</td>\n",
       "      <td>26</td>\n",
       "      <td>0</td>\n",
       "      <td>Q1</td>\n",
       "      <td>S1</td>\n",
       "      <td>9</td>\n",
       "      <td>morning</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6996</th>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>2665</td>\n",
       "      <td>AA</td>\n",
       "      <td>N196AA</td>\n",
       "      <td>2665</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>Philadelphia, PA</td>\n",
       "      <td>13303</td>\n",
       "      <td>...</td>\n",
       "      <td>2019</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>Q1</td>\n",
       "      <td>S1</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6997</th>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>5180</td>\n",
       "      <td>OH</td>\n",
       "      <td>N725PS</td>\n",
       "      <td>5180</td>\n",
       "      <td>11057</td>\n",
       "      <td>CLT</td>\n",
       "      <td>Charlotte, NC</td>\n",
       "      <td>12197</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>12</td>\n",
       "      <td>26</td>\n",
       "      <td>2</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6998</th>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>576</td>\n",
       "      <td>AA</td>\n",
       "      <td>N188US</td>\n",
       "      <td>576</td>\n",
       "      <td>14893</td>\n",
       "      <td>SMF</td>\n",
       "      <td>Sacramento, CA</td>\n",
       "      <td>11057</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>11</td>\n",
       "      <td>10</td>\n",
       "      <td>5</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>22</td>\n",
       "      <td>evening</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6999</th>\n",
       "      <td>DL</td>\n",
       "      <td>DL</td>\n",
       "      <td>4669</td>\n",
       "      <td>OO</td>\n",
       "      <td>N772SK</td>\n",
       "      <td>4669</td>\n",
       "      <td>13487</td>\n",
       "      <td>MSP</td>\n",
       "      <td>Minneapolis, MN</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>2018</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>Q2</td>\n",
       "      <td>S2</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num op_unique_carrier  \\\n",
       "6995                 AA          AA                 850                AA   \n",
       "6996                 AA          AA                2665                AA   \n",
       "6997                 AA          AA                5180                OH   \n",
       "6998                 AA          AA                 576                AA   \n",
       "6999                 DL          DL                4669                OO   \n",
       "\n",
       "     tail_num  op_carrier_fl_num  origin_airport_id origin  origin_city_name  \\\n",
       "6995   N197UW                850              13303    MIA         Miami, FL   \n",
       "6996   N196AA               2665              14100    PHL  Philadelphia, PA   \n",
       "6997   N725PS               5180              11057    CLT     Charlotte, NC   \n",
       "6998   N188US                576              14893    SMF    Sacramento, CA   \n",
       "6999   N772SK               4669              13487    MSP   Minneapolis, MN   \n",
       "\n",
       "      dest_airport_id  ...  year month  day_of_month  day_of_week quarter  \\\n",
       "6995            14100  ...  2018     3            26            0      Q1   \n",
       "6996            13303  ...  2019     2             5            1      Q1   \n",
       "6997            12197  ...  2018    12            26            2      Q4   \n",
       "6998            11057  ...  2018    11            10            5      Q4   \n",
       "6999            11193  ...  2018     5             3            3      Q2   \n",
       "\n",
       "      semester  dep_hour_of_day  dep_time_of_day  arr_hour_of_day  \\\n",
       "6995        S1                9          morning               11   \n",
       "6996        S1               10          morning               13   \n",
       "6997        S4               11          morning               13   \n",
       "6998        S4               22          evening                6   \n",
       "6999        S2                8          morning               11   \n",
       "\n",
       "      arr_time_of_day  \n",
       "6995          morning  \n",
       "6996        afternoon  \n",
       "6997        afternoon  \n",
       "6998          morning  \n",
       "6999          morning  \n",
       "\n",
       "[5 rows x 35 columns]"
      ]
     },
     "execution_count": 740,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Converting some object data types into a category data type\n",
    "\n",
    "train_flights_subset['mkt_unique_carrier'] = train_flights_subset['mkt_unique_carrier'].astype('category')\n",
    "train_flights_subset['mkt_carrier'] = train_flights_subset['mkt_carrier'].astype('category')\n",
    "train_flights_subset['op_unique_carrier'] = train_flights_subset['op_unique_carrier'].astype('category')\n",
    "train_flights_subset['tail_num'] = train_flights_subset['tail_num'].astype('category')\n",
    "train_flights_subset['origin'] = train_flights_subset['origin'].astype('category')\n",
    "train_flights_subset['dest'] = train_flights_subset['dest'].astype('category')\n",
    "train_flights_subset.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 741,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mkt_unique_carrier     category\n",
       "mkt_carrier            category\n",
       "mkt_carrier_fl_num        int64\n",
       "op_unique_carrier      category\n",
       "tail_num               category\n",
       "op_carrier_fl_num         int64\n",
       "origin_airport_id         int64\n",
       "origin                 category\n",
       "origin_city_name         object\n",
       "dest_airport_id           int64\n",
       "dest                   category\n",
       "dest_city_name           object\n",
       "arr_delay               float64\n",
       "diverted                  int64\n",
       "dup                      object\n",
       "crs_elapsed_time          int64\n",
       "actual_elapsed_time     float64\n",
       "air_time                float64\n",
       "flights                   int64\n",
       "distance                  int64\n",
       "carrier_delay           float64\n",
       "weather_delay           float64\n",
       "nas_delay               float64\n",
       "security_delay          float64\n",
       "late_aircraft_delay     float64\n",
       "year                      int64\n",
       "month                     int64\n",
       "day_of_month              int64\n",
       "day_of_week               int64\n",
       "quarter                  object\n",
       "semester                 object\n",
       "dep_hour_of_day           int64\n",
       "dep_time_of_day          object\n",
       "arr_hour_of_day           int64\n",
       "arr_time_of_day          object\n",
       "dtype: object"
      ]
     },
     "execution_count": 741,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking the data types\n",
    "\n",
    "train_flights_subset.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When we were checking the object data types, we noticed that the variables 'origin_city_name' and 'dest_city_name' have the name of the city and state separated by a comma being classified as an object. The strucutre of this information is not going to help us. So, let's separate this information in 2 more columns below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 742,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0       NY\n",
      "1       CA\n",
      "2       WV\n",
      "3       MD\n",
      "4       IN\n",
      "        ..\n",
      "6995    FL\n",
      "6996    PA\n",
      "6997    NC\n",
      "6998    CA\n",
      "6999    MN\n",
      "Name: origin_state, Length: 6975, dtype: object\n",
      "0       PA\n",
      "1       NY\n",
      "2       DC\n",
      "3       OH\n",
      "4       CO\n",
      "        ..\n",
      "6995    PA\n",
      "6996    FL\n",
      "6997    NY\n",
      "6998    NC\n",
      "6999    OH\n",
      "Name: destination_state, Length: 6975, dtype: object\n"
     ]
    }
   ],
   "source": [
    "# Separating the name of the city and state from 'origin_city_name' and 'dest_city_name'\n",
    "\n",
    "train_flights_subset['origin_state'] = train_flights_subset['origin_city_name'].apply(lambda x: x.split(', ')[1])\n",
    "train_flights_subset['destination_state'] = train_flights_subset['dest_city_name'].apply(lambda x: x.split(', ')[1])\n",
    "print(train_flights_subset['origin_state'])\n",
    "print(train_flights_subset['destination_state'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 743,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean carrier delay is 19.782575173477902\n",
      "Mean weather delay is 2.678488820354682\n",
      "Mean NAS delay is 17.033153430995117\n",
      "Mean security delay is 0.08712413261373267\n",
      "Mean late aircraft delay is 26.69082498072792\n"
     ]
    }
   ],
   "source": [
    "# As we will not have the 'dep_delay' information in the real-life, we are going to use the statistics to measure it with other delays based on the specific time of the year\n",
    "# For that, we are using the carrier_delay, weather_delay, nas_delay, security_delay, late_aircraft_delay to understand their pattern (not to use them directly in the predictions)\n",
    "\n",
    "# Analyzing first the mean of each delay type using NumPy\n",
    "\n",
    "carrier_delay_mean = np.mean(train_flights_subset['carrier_delay'])\n",
    "weather_delay_mean = np.mean(train_flights_subset['weather_delay'])\n",
    "nas_delay_mean = np.mean(train_flights_subset['nas_delay'])\n",
    "security_delay_mean = np.mean(train_flights_subset['security_delay'])\n",
    "late_aircraft_delay_mean = np.mean(train_flights_subset['late_aircraft_delay'])\n",
    "\n",
    "# Printing the results\n",
    "\n",
    "print('Mean carrier delay is', carrier_delay_mean)\n",
    "print('Mean weather delay is', weather_delay_mean)\n",
    "print('Mean NAS delay is', nas_delay_mean)\n",
    "print('Mean security delay is', security_delay_mean)\n",
    "print('Mean late aircraft delay is', late_aircraft_delay_mean)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 744,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>carrier_delay</th>\n",
       "      <th>weather_delay</th>\n",
       "      <th>nas_delay</th>\n",
       "      <th>security_delay</th>\n",
       "      <th>late_aircraft_delay</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>6975.000000</td>\n",
       "      <td>6975.000000</td>\n",
       "      <td>6975.000000</td>\n",
       "      <td>6975.000000</td>\n",
       "      <td>6975.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>25.848438</td>\n",
       "      <td>8.952115</td>\n",
       "      <td>22.161171</td>\n",
       "      <td>0.959247</td>\n",
       "      <td>26.372497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>17.033153</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1174.000000</td>\n",
       "      <td>556.000000</td>\n",
       "      <td>1272.000000</td>\n",
       "      <td>71.000000</td>\n",
       "      <td>1094.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       carrier_delay  weather_delay    nas_delay  security_delay  \\\n",
       "count    6975.000000    6975.000000  6975.000000     6975.000000   \n",
       "mean       19.782575       2.678489    17.033153        0.087124   \n",
       "std        25.848438       8.952115    22.161171        0.959247   \n",
       "min         0.000000       0.000000     0.000000        0.000000   \n",
       "25%        19.782575       2.678489    17.033153        0.087124   \n",
       "50%        19.782575       2.678489    17.033153        0.087124   \n",
       "75%        19.782575       2.678489    17.033153        0.087124   \n",
       "max      1174.000000     556.000000  1272.000000       71.000000   \n",
       "\n",
       "       late_aircraft_delay  \n",
       "count          6975.000000  \n",
       "mean             26.690825  \n",
       "std              26.372497  \n",
       "min               0.000000  \n",
       "25%              26.690825  \n",
       "50%              26.690825  \n",
       "75%              26.690825  \n",
       "max            1094.000000  "
      ]
     },
     "execution_count": 744,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculating the statistics for all delay types\n",
    "\n",
    "delay_stats = train_flights_subset[['carrier_delay', 'weather_delay', 'nas_delay', 'security_delay', 'late_aircraft_delay']].describe()\n",
    "delay_stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 745,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"8\" halign=\"left\">carrier_delay</th>\n",
       "      <th colspan=\"2\" halign=\"left\">weather_delay</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"2\" halign=\"left\">security_delay</th>\n",
       "      <th colspan=\"8\" halign=\"left\">late_aircraft_delay</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>...</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>arr_delay</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>-62.0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>...</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>1.0</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>NaN</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-54.0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>...</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>1.0</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>NaN</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-50.0</th>\n",
       "      <td>3.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>...</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>3.0</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>4.351168e-15</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-49.0</th>\n",
       "      <td>2.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>...</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>2.0</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-47.0</th>\n",
       "      <td>2.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>19.782575</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.678489</td>\n",
       "      <td>...</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>0.087124</td>\n",
       "      <td>2.0</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "      <td>26.690825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>836.0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>955.0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>878.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>878.000000</td>\n",
       "      <td>878.000000</td>\n",
       "      <td>878.000000</td>\n",
       "      <td>878.000000</td>\n",
       "      <td>878.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>68.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>68.000000</td>\n",
       "      <td>68.000000</td>\n",
       "      <td>68.000000</td>\n",
       "      <td>68.000000</td>\n",
       "      <td>68.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1111.0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1094.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1094.000000</td>\n",
       "      <td>1094.000000</td>\n",
       "      <td>1094.000000</td>\n",
       "      <td>1094.000000</td>\n",
       "      <td>1094.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1174.0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1174.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1174.000000</td>\n",
       "      <td>1174.000000</td>\n",
       "      <td>1174.000000</td>\n",
       "      <td>1174.000000</td>\n",
       "      <td>1174.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1272.0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>276 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          carrier_delay                                              \\\n",
       "                  count         mean  std          min          25%   \n",
       "arr_delay                                                             \n",
       "-62.0               1.0    19.782575  NaN    19.782575    19.782575   \n",
       "-54.0               1.0    19.782575  NaN    19.782575    19.782575   \n",
       "-50.0               3.0    19.782575  0.0    19.782575    19.782575   \n",
       "-49.0               2.0    19.782575  0.0    19.782575    19.782575   \n",
       "-47.0               2.0    19.782575  0.0    19.782575    19.782575   \n",
       "...                 ...          ...  ...          ...          ...   \n",
       " 836.0              1.0     0.000000  NaN     0.000000     0.000000   \n",
       " 955.0              1.0   878.000000  NaN   878.000000   878.000000   \n",
       " 1111.0             1.0    17.000000  NaN    17.000000    17.000000   \n",
       " 1174.0             1.0  1174.000000  NaN  1174.000000  1174.000000   \n",
       " 1272.0             1.0     0.000000  NaN     0.000000     0.000000   \n",
       "\n",
       "                                                 weather_delay            ...  \\\n",
       "                   50%          75%          max         count      mean  ...   \n",
       "arr_delay                                                                 ...   \n",
       "-62.0        19.782575    19.782575    19.782575           1.0  2.678489  ...   \n",
       "-54.0        19.782575    19.782575    19.782575           1.0  2.678489  ...   \n",
       "-50.0        19.782575    19.782575    19.782575           3.0  2.678489  ...   \n",
       "-49.0        19.782575    19.782575    19.782575           2.0  2.678489  ...   \n",
       "-47.0        19.782575    19.782575    19.782575           2.0  2.678489  ...   \n",
       "...                ...          ...          ...           ...       ...  ...   \n",
       " 836.0        0.000000     0.000000     0.000000           1.0  0.000000  ...   \n",
       " 955.0      878.000000   878.000000   878.000000           1.0  0.000000  ...   \n",
       " 1111.0      17.000000    17.000000    17.000000           1.0  0.000000  ...   \n",
       " 1174.0    1174.000000  1174.000000  1174.000000           1.0  0.000000  ...   \n",
       " 1272.0       0.000000     0.000000     0.000000           1.0  0.000000  ...   \n",
       "\n",
       "          security_delay           late_aircraft_delay               \\\n",
       "                     75%       max               count         mean   \n",
       "arr_delay                                                             \n",
       "-62.0           0.087124  0.087124                 1.0    26.690825   \n",
       "-54.0           0.087124  0.087124                 1.0    26.690825   \n",
       "-50.0           0.087124  0.087124                 3.0    26.690825   \n",
       "-49.0           0.087124  0.087124                 2.0    26.690825   \n",
       "-47.0           0.087124  0.087124                 2.0    26.690825   \n",
       "...                  ...       ...                 ...          ...   \n",
       " 836.0          0.000000  0.000000                 1.0     0.000000   \n",
       " 955.0          0.000000  0.000000                 1.0    68.000000   \n",
       " 1111.0         0.000000  0.000000                 1.0  1094.000000   \n",
       " 1174.0         0.000000  0.000000                 1.0     0.000000   \n",
       " 1272.0         0.000000  0.000000                 1.0     0.000000   \n",
       "\n",
       "                                                                             \\\n",
       "                    std          min          25%          50%          75%   \n",
       "arr_delay                                                                     \n",
       "-62.0               NaN    26.690825    26.690825    26.690825    26.690825   \n",
       "-54.0               NaN    26.690825    26.690825    26.690825    26.690825   \n",
       "-50.0      4.351168e-15    26.690825    26.690825    26.690825    26.690825   \n",
       "-49.0      0.000000e+00    26.690825    26.690825    26.690825    26.690825   \n",
       "-47.0      0.000000e+00    26.690825    26.690825    26.690825    26.690825   \n",
       "...                 ...          ...          ...          ...          ...   \n",
       " 836.0              NaN     0.000000     0.000000     0.000000     0.000000   \n",
       " 955.0              NaN    68.000000    68.000000    68.000000    68.000000   \n",
       " 1111.0             NaN  1094.000000  1094.000000  1094.000000  1094.000000   \n",
       " 1174.0             NaN     0.000000     0.000000     0.000000     0.000000   \n",
       " 1272.0             NaN     0.000000     0.000000     0.000000     0.000000   \n",
       "\n",
       "                        \n",
       "                   max  \n",
       "arr_delay               \n",
       "-62.0        26.690825  \n",
       "-54.0        26.690825  \n",
       "-50.0        26.690825  \n",
       "-49.0        26.690825  \n",
       "-47.0        26.690825  \n",
       "...                ...  \n",
       " 836.0        0.000000  \n",
       " 955.0       68.000000  \n",
       " 1111.0    1094.000000  \n",
       " 1174.0       0.000000  \n",
       " 1272.0       0.000000  \n",
       "\n",
       "[276 rows x 40 columns]"
      ]
     },
     "execution_count": 745,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Now the statistics in regards of our target variable (arr_delay)\n",
    "\n",
    "delay_stats = train_flights_subset.groupby('arr_delay')[['carrier_delay', 'weather_delay', 'nas_delay', 'security_delay', 'late_aircraft_delay']].describe()\n",
    "delay_stats"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From this statistics, it is noted that the max count of delays are caused by the variables 'nas_delay', 'carrier_delay' and 'late_aircraft_delay'. Also, these same variables are among the ones with the highest mean of the delays, so this might indicate that most of the factors that cause flight delays are related to airplanes traffic, maintanance and service.\n",
    "\n",
    "From that, let's try to calculate the mean of the arrival delays considering the 'mkt_unique_carrier' variable to check the count of delays per airline."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 746,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>...</th>\n",
       "      <th>quarter</th>\n",
       "      <th>semester</th>\n",
       "      <th>dep_hour_of_day</th>\n",
       "      <th>dep_time_of_day</th>\n",
       "      <th>arr_hour_of_day</th>\n",
       "      <th>arr_time_of_day</th>\n",
       "      <th>origin_state</th>\n",
       "      <th>destination_state</th>\n",
       "      <th>arr_delay_mean_by_mkt_unique_carrier</th>\n",
       "      <th>arr_delay_mean_by_op_unique_carrier</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>14</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>NY</td>\n",
       "      <td>PA</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.953181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>12478</td>\n",
       "      <td>...</td>\n",
       "      <td>Q4</td>\n",
       "      <td>S4</td>\n",
       "      <td>23</td>\n",
       "      <td>evening</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>CA</td>\n",
       "      <td>NY</td>\n",
       "      <td>6.360829</td>\n",
       "      <td>6.360829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>12264</td>\n",
       "      <td>...</td>\n",
       "      <td>Q3</td>\n",
       "      <td>S3</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "      <td>WV</td>\n",
       "      <td>DC</td>\n",
       "      <td>10.231880</td>\n",
       "      <td>7.331876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>Q1</td>\n",
       "      <td>S1</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "      <td>MD</td>\n",
       "      <td>OH</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>11292</td>\n",
       "      <td>...</td>\n",
       "      <td>Q2</td>\n",
       "      <td>S2</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "      <td>IN</td>\n",
       "      <td>CO</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 39 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num op_unique_carrier  \\\n",
       "0                 AA          AA                4850                PT   \n",
       "1                 B6          B6                 168                B6   \n",
       "2                 UA          UA                5068                OO   \n",
       "3                 WN          WN                 388                WN   \n",
       "4                 WN          WN                1353                WN   \n",
       "\n",
       "  tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0   N657AE               4850              15096    SYR   \n",
       "1   N633JB                168              13796    OAK   \n",
       "2   N975SW               5068              11027    CKB   \n",
       "3   N968WN                388              10821    BWI   \n",
       "4   N8575Z               1353              12339    IND   \n",
       "\n",
       "          origin_city_name  dest_airport_id  ... quarter semester  \\\n",
       "0             Syracuse, NY            14100  ...      Q4       S4   \n",
       "1              Oakland, CA            12478  ...      Q4       S4   \n",
       "2  Clarksburg/Fairmont, WV            12264  ...      Q3       S3   \n",
       "3            Baltimore, MD            11193  ...      Q1       S1   \n",
       "4         Indianapolis, IN            11292  ...      Q2       S2   \n",
       "\n",
       "   dep_hour_of_day  dep_time_of_day arr_hour_of_day  arr_time_of_day  \\\n",
       "0               13        afternoon              14        afternoon   \n",
       "1               23          evening               8          morning   \n",
       "2                6          morning               7          morning   \n",
       "3                8          morning              10          morning   \n",
       "4                6          morning               7          morning   \n",
       "\n",
       "   origin_state  destination_state  arr_delay_mean_by_mkt_unique_carrier  \\\n",
       "0            NY                 PA                              5.328596   \n",
       "1            CA                 NY                              6.360829   \n",
       "2            WV                 DC                             10.231880   \n",
       "3            MD                 OH                              2.438350   \n",
       "4            IN                 CO                              2.438350   \n",
       "\n",
       "   arr_delay_mean_by_op_unique_carrier  \n",
       "0                             5.953181  \n",
       "1                             6.360829  \n",
       "2                             7.331876  \n",
       "3                             2.438350  \n",
       "4                             2.438350  \n",
       "\n",
       "[5 rows x 39 columns]"
      ]
     },
     "execution_count": 746,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculating the mean of the 'arr_delay' per 'mkt_unique_carrier' and 'op_unique_carrier'\n",
    "\n",
    "# To analyze flight delays by the airline that sold the ticket, we are using the 'mkt_unique_carrier' lines that marketed and sold the tickets\n",
    "\n",
    "train_flights_subset['arr_delay_mean_by_mkt_unique_carrier']=train_flights_subset.groupby(['mkt_unique_carrier'])['arr_delay'].transform(np.mean)\n",
    "\n",
    "# To analyze delays by the airline that physically operated the flight, we are using the 'op_unique_carrier' \n",
    "\n",
    "train_flights_subset['arr_delay_mean_by_op_unique_carrier']=train_flights_subset.groupby(['op_unique_carrier'])['arr_delay'].transform(np.mean)\n",
    "train_flights_subset.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 747,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "     mkt_unique_carrier  arr_delay_mean_by_mkt_unique_carrier\n",
      "4028                 UA                              10.23188\n",
      "5623                 UA                              10.23188\n",
      "1294                 UA                              10.23188\n",
      "5614                 UA                              10.23188\n",
      "1291                 UA                              10.23188\n",
      "1290                 UA                              10.23188\n",
      "6178                 UA                              10.23188\n",
      "5945                 UA                              10.23188\n",
      "4120                 UA                              10.23188\n",
      "4121                 UA                              10.23188\n"
     ]
    }
   ],
   "source": [
    "# Sorting by arr_delay_mean_by_mkt_unique_carrier\n",
    "\n",
    "train_flights_subset_sorted = train_flights_subset.sort_values('arr_delay_mean_by_mkt_unique_carrier', ascending=False)\n",
    "print(train_flights_subset_sorted[['mkt_unique_carrier', 'arr_delay_mean_by_mkt_unique_carrier']].head(10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 748,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  op_unique_carrier  arr_delay_mean_by_op_unique_carrier\n",
      "0                PT                             5.953181\n",
      "1                B6                             6.360829\n",
      "2                OO                             7.331876\n",
      "3                WN                             2.438350\n",
      "4                WN                             2.438350\n",
      "5                MQ                             5.849723\n",
      "6                DL                            -0.301824\n",
      "7                WN                             2.438350\n",
      "8                AA                             5.396410\n",
      "9                AA                             5.396410\n"
     ]
    }
   ],
   "source": [
    "# Sorting by arr_delay_mean_by_op_unique_carrier\n",
    "\n",
    "train_flights_subset.sort_values('arr_delay_mean_by_op_unique_carrier', ascending=False)\n",
    "print(train_flights_subset[['op_unique_carrier', 'arr_delay_mean_by_op_unique_carrier']].head(10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is noted that from the 'mkt_unique_carrier' (marketing unique carrier) variable, which refers to the airline that sold the ticket for the flight, the United Airlines (UA) has the highest mean of arrival delay in the results. \n",
    "\n",
    "In regards to the 'op_unique_carrier', the SkyWest Airlines (OO), which is a regional airline that operates flights on behalf of other airlines, JetBlue Airways (B6), which is a low-cost airline based in the United States, and Piedmont Airlines (PT), which is a regional airline that operates flights on behalf of American Airlines, are among with the higher arrival delay in the group of the airlines that physically operate the flights.\n",
    "\n",
    "Therefore, the results indicate that *United Airlines* has the highest mean of arrival delay among all airlines, while *SkyWest Airlines*, *JetBlue Airways*, and *Piedmont Airlines* have a high mean of arrival delays when they are the airlines physically operating the flight."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 749,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the distribution of the 'arr_delay' to visualize it before encoding it\n",
    "\n",
    "sns.histplot(data=train_flights_subset, x='arr_delay', bins=50, color='blue', binwidth=10)\n",
    "plt.xticks(fontsize=10)\n",
    "plt.yticks(fontsize=10)\n",
    "plt.xlabel('Arrival delay (minutes)', fontsize=10)\n",
    "plt.ylabel('Count', fontsize=10)\n",
    "plt.title('Distribution of Arrival Delay', fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 750,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>...</th>\n",
       "      <th>semester</th>\n",
       "      <th>dep_hour_of_day</th>\n",
       "      <th>dep_time_of_day</th>\n",
       "      <th>arr_hour_of_day</th>\n",
       "      <th>arr_time_of_day</th>\n",
       "      <th>origin_state</th>\n",
       "      <th>destination_state</th>\n",
       "      <th>arr_delay_mean_by_mkt_unique_carrier</th>\n",
       "      <th>arr_delay_mean_by_op_unique_carrier</th>\n",
       "      <th>arr_delay_encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>4850</td>\n",
       "      <td>PT</td>\n",
       "      <td>N657AE</td>\n",
       "      <td>4850</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>S4</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>14</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>NY</td>\n",
       "      <td>PA</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.953181</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B6</td>\n",
       "      <td>B6</td>\n",
       "      <td>168</td>\n",
       "      <td>B6</td>\n",
       "      <td>N633JB</td>\n",
       "      <td>168</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>12478</td>\n",
       "      <td>...</td>\n",
       "      <td>S4</td>\n",
       "      <td>23</td>\n",
       "      <td>evening</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>CA</td>\n",
       "      <td>NY</td>\n",
       "      <td>6.360829</td>\n",
       "      <td>6.360829</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>5068</td>\n",
       "      <td>OO</td>\n",
       "      <td>N975SW</td>\n",
       "      <td>5068</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>12264</td>\n",
       "      <td>...</td>\n",
       "      <td>S3</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "      <td>WV</td>\n",
       "      <td>DC</td>\n",
       "      <td>10.231880</td>\n",
       "      <td>7.331876</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>388</td>\n",
       "      <td>WN</td>\n",
       "      <td>N968WN</td>\n",
       "      <td>388</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>S1</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "      <td>MD</td>\n",
       "      <td>OH</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WN</td>\n",
       "      <td>WN</td>\n",
       "      <td>1353</td>\n",
       "      <td>WN</td>\n",
       "      <td>N8575Z</td>\n",
       "      <td>1353</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>11292</td>\n",
       "      <td>...</td>\n",
       "      <td>S2</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "      <td>IN</td>\n",
       "      <td>CO</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  mkt_unique_carrier mkt_carrier  mkt_carrier_fl_num op_unique_carrier  \\\n",
       "0                 AA          AA                4850                PT   \n",
       "1                 B6          B6                 168                B6   \n",
       "2                 UA          UA                5068                OO   \n",
       "3                 WN          WN                 388                WN   \n",
       "4                 WN          WN                1353                WN   \n",
       "\n",
       "  tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0   N657AE               4850              15096    SYR   \n",
       "1   N633JB                168              13796    OAK   \n",
       "2   N975SW               5068              11027    CKB   \n",
       "3   N968WN                388              10821    BWI   \n",
       "4   N8575Z               1353              12339    IND   \n",
       "\n",
       "          origin_city_name  dest_airport_id  ... semester dep_hour_of_day  \\\n",
       "0             Syracuse, NY            14100  ...       S4              13   \n",
       "1              Oakland, CA            12478  ...       S4              23   \n",
       "2  Clarksburg/Fairmont, WV            12264  ...       S3               6   \n",
       "3            Baltimore, MD            11193  ...       S1               8   \n",
       "4         Indianapolis, IN            11292  ...       S2               6   \n",
       "\n",
       "   dep_time_of_day  arr_hour_of_day arr_time_of_day  origin_state  \\\n",
       "0        afternoon               14       afternoon            NY   \n",
       "1          evening                8         morning            CA   \n",
       "2          morning                7         morning            WV   \n",
       "3          morning               10         morning            MD   \n",
       "4          morning                7         morning            IN   \n",
       "\n",
       "   destination_state  arr_delay_mean_by_mkt_unique_carrier  \\\n",
       "0                 PA                              5.328596   \n",
       "1                 NY                              6.360829   \n",
       "2                 DC                             10.231880   \n",
       "3                 OH                              2.438350   \n",
       "4                 CO                              2.438350   \n",
       "\n",
       "   arr_delay_mean_by_op_unique_carrier  arr_delay_encoded  \n",
       "0                             5.953181                  0  \n",
       "1                             6.360829                  1  \n",
       "2                             7.331876                  0  \n",
       "3                             2.438350                  1  \n",
       "4                             2.438350                  1  \n",
       "\n",
       "[5 rows x 40 columns]"
      ]
     },
     "execution_count": 750,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use the encoding technique transform the 'arr_delay' variable in a binary one with a new colum\n",
    "\n",
    "arr_delay_encoded = []\n",
    "for delay in train_flights_subset['arr_delay']:\n",
    "    if delay > 0:\n",
    "        arr_delay_encoded.append(1)\n",
    "    else:\n",
    "        arr_delay_encoded.append(0)\n",
    "\n",
    "train_flights_subset['arr_delay_encoded'] = arr_delay_encoded\n",
    "train_flights_subset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 751,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/base_env/lib/python3.8/site-packages/sklearn/preprocessing/_label.py:116: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>...</th>\n",
       "      <th>semester</th>\n",
       "      <th>dep_hour_of_day</th>\n",
       "      <th>dep_time_of_day</th>\n",
       "      <th>arr_hour_of_day</th>\n",
       "      <th>arr_time_of_day</th>\n",
       "      <th>origin_state</th>\n",
       "      <th>destination_state</th>\n",
       "      <th>arr_delay_mean_by_mkt_unique_carrier</th>\n",
       "      <th>arr_delay_mean_by_op_unique_carrier</th>\n",
       "      <th>arr_delay_encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6990</th>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>1852</td>\n",
       "      <td>7</td>\n",
       "      <td>3875</td>\n",
       "      <td>1852</td>\n",
       "      <td>10397</td>\n",
       "      <td>ATL</td>\n",
       "      <td>Atlanta, GA</td>\n",
       "      <td>14027</td>\n",
       "      <td>...</td>\n",
       "      <td>S2</td>\n",
       "      <td>15</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>17</td>\n",
       "      <td>evening</td>\n",
       "      <td>GA</td>\n",
       "      <td>FL</td>\n",
       "      <td>2.144038</td>\n",
       "      <td>-0.301824</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6991</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2419</td>\n",
       "      <td>15</td>\n",
       "      <td>691</td>\n",
       "      <td>2419</td>\n",
       "      <td>10781</td>\n",
       "      <td>BTR</td>\n",
       "      <td>Baton Rouge, LA</td>\n",
       "      <td>11298</td>\n",
       "      <td>...</td>\n",
       "      <td>S4</td>\n",
       "      <td>12</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>14</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>LA</td>\n",
       "      <td>TX</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.849723</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6992</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>388</td>\n",
       "      <td>2</td>\n",
       "      <td>1675</td>\n",
       "      <td>388</td>\n",
       "      <td>14908</td>\n",
       "      <td>SNA</td>\n",
       "      <td>Santa Ana, CA</td>\n",
       "      <td>14747</td>\n",
       "      <td>...</td>\n",
       "      <td>S1</td>\n",
       "      <td>15</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>18</td>\n",
       "      <td>evening</td>\n",
       "      <td>CA</td>\n",
       "      <td>WA</td>\n",
       "      <td>1.305767</td>\n",
       "      <td>1.322588</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6993</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>388</td>\n",
       "      <td>23</td>\n",
       "      <td>2264</td>\n",
       "      <td>388</td>\n",
       "      <td>14570</td>\n",
       "      <td>RNO</td>\n",
       "      <td>Reno, NV</td>\n",
       "      <td>14831</td>\n",
       "      <td>...</td>\n",
       "      <td>S1</td>\n",
       "      <td>18</td>\n",
       "      <td>evening</td>\n",
       "      <td>19</td>\n",
       "      <td>evening</td>\n",
       "      <td>NV</td>\n",
       "      <td>CA</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6994</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>374</td>\n",
       "      <td>23</td>\n",
       "      <td>2976</td>\n",
       "      <td>374</td>\n",
       "      <td>14679</td>\n",
       "      <td>SAN</td>\n",
       "      <td>San Diego, CA</td>\n",
       "      <td>10821</td>\n",
       "      <td>...</td>\n",
       "      <td>S1</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>14</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>CA</td>\n",
       "      <td>MD</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6995</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>660</td>\n",
       "      <td>1</td>\n",
       "      <td>443</td>\n",
       "      <td>660</td>\n",
       "      <td>13303</td>\n",
       "      <td>MIA</td>\n",
       "      <td>Miami, FL</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>S1</td>\n",
       "      <td>9</td>\n",
       "      <td>morning</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "      <td>FL</td>\n",
       "      <td>PA</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.396410</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6996</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1984</td>\n",
       "      <td>1</td>\n",
       "      <td>438</td>\n",
       "      <td>1984</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>Philadelphia, PA</td>\n",
       "      <td>13303</td>\n",
       "      <td>...</td>\n",
       "      <td>S1</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>PA</td>\n",
       "      <td>FL</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.396410</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6997</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3337</td>\n",
       "      <td>17</td>\n",
       "      <td>2350</td>\n",
       "      <td>3337</td>\n",
       "      <td>11057</td>\n",
       "      <td>CLT</td>\n",
       "      <td>Charlotte, NC</td>\n",
       "      <td>12197</td>\n",
       "      <td>...</td>\n",
       "      <td>S4</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>NC</td>\n",
       "      <td>NY</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>6.229542</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6998</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>445</td>\n",
       "      <td>1</td>\n",
       "      <td>414</td>\n",
       "      <td>445</td>\n",
       "      <td>14893</td>\n",
       "      <td>SMF</td>\n",
       "      <td>Sacramento, CA</td>\n",
       "      <td>11057</td>\n",
       "      <td>...</td>\n",
       "      <td>S4</td>\n",
       "      <td>22</td>\n",
       "      <td>evening</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>CA</td>\n",
       "      <td>NC</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.396410</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6999</th>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3050</td>\n",
       "      <td>18</td>\n",
       "      <td>2532</td>\n",
       "      <td>3050</td>\n",
       "      <td>13487</td>\n",
       "      <td>MSP</td>\n",
       "      <td>Minneapolis, MN</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>S2</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "      <td>MN</td>\n",
       "      <td>OH</td>\n",
       "      <td>2.144038</td>\n",
       "      <td>7.331876</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
       "6990                   3            3                1852                  7   \n",
       "6991                   0            0                2419                 15   \n",
       "6992                   1            1                 388                  2   \n",
       "6993                  10           10                 388                 23   \n",
       "6994                  10           10                 374                 23   \n",
       "6995                   0            0                 660                  1   \n",
       "6996                   0            0                1984                  1   \n",
       "6997                   0            0                3337                 17   \n",
       "6998                   0            0                 445                  1   \n",
       "6999                   3            3                3050                 18   \n",
       "\n",
       "      tail_num  op_carrier_fl_num  origin_airport_id origin  origin_city_name  \\\n",
       "6990      3875               1852              10397    ATL       Atlanta, GA   \n",
       "6991       691               2419              10781    BTR   Baton Rouge, LA   \n",
       "6992      1675                388              14908    SNA     Santa Ana, CA   \n",
       "6993      2264                388              14570    RNO          Reno, NV   \n",
       "6994      2976                374              14679    SAN     San Diego, CA   \n",
       "6995       443                660              13303    MIA         Miami, FL   \n",
       "6996       438               1984              14100    PHL  Philadelphia, PA   \n",
       "6997      2350               3337              11057    CLT     Charlotte, NC   \n",
       "6998       414                445              14893    SMF    Sacramento, CA   \n",
       "6999      2532               3050              13487    MSP   Minneapolis, MN   \n",
       "\n",
       "      dest_airport_id  ... semester dep_hour_of_day  dep_time_of_day  \\\n",
       "6990            14027  ...       S2              15        afternoon   \n",
       "6991            11298  ...       S4              12        afternoon   \n",
       "6992            14747  ...       S1              15        afternoon   \n",
       "6993            14831  ...       S1              18          evening   \n",
       "6994            10821  ...       S1               6          morning   \n",
       "6995            14100  ...       S1               9          morning   \n",
       "6996            13303  ...       S1              10          morning   \n",
       "6997            12197  ...       S4              11          morning   \n",
       "6998            11057  ...       S4              22          evening   \n",
       "6999            11193  ...       S2               8          morning   \n",
       "\n",
       "      arr_hour_of_day arr_time_of_day  origin_state  destination_state  \\\n",
       "6990               17         evening            GA                 FL   \n",
       "6991               14       afternoon            LA                 TX   \n",
       "6992               18         evening            CA                 WA   \n",
       "6993               19         evening            NV                 CA   \n",
       "6994               14       afternoon            CA                 MD   \n",
       "6995               11         morning            FL                 PA   \n",
       "6996               13       afternoon            PA                 FL   \n",
       "6997               13       afternoon            NC                 NY   \n",
       "6998                6         morning            CA                 NC   \n",
       "6999               11         morning            MN                 OH   \n",
       "\n",
       "      arr_delay_mean_by_mkt_unique_carrier  \\\n",
       "6990                              2.144038   \n",
       "6991                              5.328596   \n",
       "6992                              1.305767   \n",
       "6993                              2.438350   \n",
       "6994                              2.438350   \n",
       "6995                              5.328596   \n",
       "6996                              5.328596   \n",
       "6997                              5.328596   \n",
       "6998                              5.328596   \n",
       "6999                              2.144038   \n",
       "\n",
       "      arr_delay_mean_by_op_unique_carrier  arr_delay_encoded  \n",
       "6990                            -0.301824                  0  \n",
       "6991                             5.849723                  0  \n",
       "6992                             1.322588                  1  \n",
       "6993                             2.438350                  0  \n",
       "6994                             2.438350                  0  \n",
       "6995                             5.396410                  0  \n",
       "6996                             5.396410                  0  \n",
       "6997                             6.229542                  0  \n",
       "6998                             5.396410                  0  \n",
       "6999                             7.331876                  0  \n",
       "\n",
       "[10 rows x 40 columns]"
      ]
     },
     "execution_count": 751,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Applying label encoding to our categorical variables\n",
    "# We are going to use label encoding even though we know that this is better used for variables with some order\n",
    "# We haven't used one-hot encoding for that because this would increase the size of the dataset as we have lots of variables to convert\n",
    "\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "label_encoder = LabelEncoder()\n",
    "\n",
    "train_flights_subset['mkt_unique_carrier'] = label_encoder.fit_transform(train_flights_subset[['mkt_unique_carrier']])\n",
    "train_flights_subset['mkt_carrier'] = label_encoder.fit_transform(train_flights_subset[['mkt_carrier']])\n",
    "train_flights_subset['mkt_carrier_fl_num'] = label_encoder.fit_transform(train_flights_subset[['mkt_carrier_fl_num']])\n",
    "train_flights_subset['op_unique_carrier'] = label_encoder.fit_transform(train_flights_subset[['op_unique_carrier']])\n",
    "train_flights_subset['op_carrier_fl_num'] = label_encoder.fit_transform(train_flights_subset[['op_carrier_fl_num']])\n",
    "train_flights_subset['tail_num'] = label_encoder.fit_transform(train_flights_subset[['tail_num']])\n",
    "train_flights_subset.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 752,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating a function to calculate the arrival delay with the variables 'origin_airport_id' and 'dest_airport_id'\n",
    "\n",
    "def total_delay_by_airport_pair(df, origin_airport_id, dest_airport_id):\n",
    "    \"\"\"\n",
    "    This function calculates the total arrival delay for a specific pair of origin and destination airports variables.\n",
    "    These are the arguments or inputs of the function:\n",
    "        df: a dataframe with specific columns \"origin_airport_id\", \"dest_airport_id\", and \"arr_delay\";\n",
    "        origin_airport_id: an integer representing the ID of the origin airport\n",
    "        dest_airport_id: an integer representing the ID of the destination airport\n",
    "    The output of this function is:\n",
    "        The total arrival delay for the given airport pair of ids\n",
    "    \"\"\"\n",
    "    total_delay = 0\n",
    "    total_delay = df.loc[(df['origin_airport_id'] == origin_airport_id) & (df['dest_airport_id'] == dest_airport_id), 'arr_delay'].sum()\n",
    "    return total_delay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 753,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8.0\n"
     ]
    }
   ],
   "source": [
    "# Testing the function\n",
    "\n",
    "total_delay = total_delay_by_airport_pair(train_flights_subset, 10397, 14027)\n",
    "print(total_delay)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 754,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The total delay of the airport 11193 to the airport 13930 is 1231.0.\n",
      "The total delay of the airport 11618 to the airport 13487 is 1174.0.\n",
      "The total delay of the airport 11540 to the airport 12892 is 1111.0.\n"
     ]
    }
   ],
   "source": [
    "# Calculating the total airport traffic with the variables 'origin_airport_id' and 'dest_airport_id'\n",
    "\n",
    "total_delay = total_delay_by_airport_pair(train_flights_subset, 11193, 13930)\n",
    "print(f'The total delay of the airport 11193 to the airport 13930 is {total_delay}.')\n",
    "total_delay = total_delay_by_airport_pair(train_flights_subset, 11618, 13487)\n",
    "print(f'The total delay of the airport 11618 to the airport 13487 is {total_delay}.')\n",
    "total_delay = total_delay_by_airport_pair(train_flights_subset, 11540, 12892)\n",
    "print(f'The total delay of the airport 11540 to the airport 12892 is {total_delay}.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 755,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "     mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
      "254                   0            0                2324                 15   \n",
      "\n",
      "     tail_num  op_carrier_fl_num  origin_airport_id origin origin_city_name  \\\n",
      "254      3434               2324              11193    CVG   Cincinnati, OH   \n",
      "\n",
      "     dest_airport_id  ... semester dep_hour_of_day  dep_time_of_day  \\\n",
      "254            13930  ...       S1               9          morning   \n",
      "\n",
      "     arr_hour_of_day arr_time_of_day  origin_state  destination_state  \\\n",
      "254               10         morning            OH                 IL   \n",
      "\n",
      "     arr_delay_mean_by_mkt_unique_carrier  \\\n",
      "254                              5.328596   \n",
      "\n",
      "     arr_delay_mean_by_op_unique_carrier  arr_delay_encoded  \n",
      "254                             5.849723                  0  \n",
      "\n",
      "[1 rows x 40 columns]\n"
     ]
    }
   ],
   "source": [
    "# Filtering the df to check the name of the airports id\n",
    "\n",
    "filtered_train_flights_subset = train_flights_subset[(train_flights_subset['origin_airport_id'] == 11193) & (train_flights_subset['dest_airport_id'] == 13930)]\n",
    "print(filtered_train_flights_subset.head(1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 756,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
      "5468                   3            3                2611                 18   \n",
      "\n",
      "      tail_num  op_carrier_fl_num  origin_airport_id origin origin_city_name  \\\n",
      "5468      2749               2611              11618    EWR       Newark, NJ   \n",
      "\n",
      "      dest_airport_id  ... semester dep_hour_of_day  dep_time_of_day  \\\n",
      "5468            13487  ...       S1              19          evening   \n",
      "\n",
      "      arr_hour_of_day arr_time_of_day  origin_state  destination_state  \\\n",
      "5468               21         evening            NJ                 MN   \n",
      "\n",
      "      arr_delay_mean_by_mkt_unique_carrier  \\\n",
      "5468                              2.144038   \n",
      "\n",
      "      arr_delay_mean_by_op_unique_carrier  arr_delay_encoded  \n",
      "5468                             7.331876                  1  \n",
      "\n",
      "[1 rows x 40 columns]\n"
     ]
    }
   ],
   "source": [
    "# Filtering the df to check the name of the airports id\n",
    "\n",
    "filtered_train_flights_subset = train_flights_subset[(train_flights_subset['origin_airport_id'] == 11618) & (train_flights_subset['dest_airport_id'] == 13487)]\n",
    "print(filtered_train_flights_subset.head(1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 757,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
      "1727                   0            0                3871                  6   \n",
      "\n",
      "      tail_num  op_carrier_fl_num  origin_airport_id origin origin_city_name  \\\n",
      "1727       499               3871              11540    ELP      El Paso, TX   \n",
      "\n",
      "      dest_airport_id  ... semester dep_hour_of_day  dep_time_of_day  \\\n",
      "1727            12892  ...       S2               6          morning   \n",
      "\n",
      "      arr_hour_of_day arr_time_of_day  origin_state  destination_state  \\\n",
      "1727                7         morning            TX                 CA   \n",
      "\n",
      "      arr_delay_mean_by_mkt_unique_carrier  \\\n",
      "1727                              5.328596   \n",
      "\n",
      "      arr_delay_mean_by_op_unique_carrier  arr_delay_encoded  \n",
      "1727                            17.384559                  1  \n",
      "\n",
      "[1 rows x 40 columns]\n"
     ]
    }
   ],
   "source": [
    "# Filtering the df to check the name of the airports id\n",
    "\n",
    "filtered_train_flights_subset = train_flights_subset[(train_flights_subset['origin_airport_id'] == 11540) & (train_flights_subset['dest_airport_id'] == 12892)]\n",
    "print(filtered_train_flights_subset.head(1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From those results, the airport trajectories with origin in (1) Cincinnati (CVG) to Illinois (ORD), (2) Newark (EWR) to Minneapolis (MSP) and (3) El Paso (ELP) to Los Angeles (LAX) are the ones with the higher counts of total arrival delays."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 758,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating a function to calculate the airport traffic with the variables 'origin' and 'dest'\n",
    "\n",
    "def count_airport_traffic(df):\n",
    "    \"\"\"\n",
    "    This function counts the number of flights for each origin-destination airport pair.\n",
    "    The inputs or arguments are:\n",
    "        df: a data frame containing the columns \"origin_airport_id\" and \"dest_airport_id\"\n",
    "    The outputs of this function are:\n",
    "        A new pandas df with two columns: \"origin_dest_pair\" (a string representing the origin-destination airport pair)\n",
    "        and \"flight_count\" (an integer representing the number of flights for that airport pair)\n",
    "    \"\"\"\n",
    "    df['origin_dest_pair'] = df['origin'] + '-' + df['dest']\n",
    "    airport_traffic = df.groupby('origin_dest_pair').size().reset_index(name='flight_count')\n",
    "    return airport_traffic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 759,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mkt_unique_carrier                        int64\n",
       "mkt_carrier                               int64\n",
       "mkt_carrier_fl_num                        int64\n",
       "op_unique_carrier                         int64\n",
       "tail_num                                  int64\n",
       "op_carrier_fl_num                         int64\n",
       "origin_airport_id                         int64\n",
       "origin                                   object\n",
       "origin_city_name                         object\n",
       "dest_airport_id                           int64\n",
       "dest                                     object\n",
       "dest_city_name                           object\n",
       "arr_delay                               float64\n",
       "diverted                                  int64\n",
       "dup                                      object\n",
       "crs_elapsed_time                          int64\n",
       "actual_elapsed_time                     float64\n",
       "air_time                                float64\n",
       "flights                                   int64\n",
       "distance                                  int64\n",
       "carrier_delay                           float64\n",
       "weather_delay                           float64\n",
       "nas_delay                               float64\n",
       "security_delay                          float64\n",
       "late_aircraft_delay                     float64\n",
       "year                                      int64\n",
       "month                                     int64\n",
       "day_of_month                              int64\n",
       "day_of_week                               int64\n",
       "quarter                                  object\n",
       "semester                                 object\n",
       "dep_hour_of_day                           int64\n",
       "dep_time_of_day                          object\n",
       "arr_hour_of_day                           int64\n",
       "arr_time_of_day                          object\n",
       "origin_state                             object\n",
       "destination_state                        object\n",
       "arr_delay_mean_by_mkt_unique_carrier    float64\n",
       "arr_delay_mean_by_op_unique_carrier     float64\n",
       "arr_delay_encoded                         int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 759,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Converting the \"origin\" and \"dest\" columns to strings before testing the function\n",
    "\n",
    "train_flights_subset[\"origin\"] = train_flights_subset[\"origin\"].astype(str)\n",
    "train_flights_subset[\"dest\"] = train_flights_subset[\"dest\"].astype(str)\n",
    "train_flights_subset.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 760,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "     origin_dest_pair  flight_count\n",
      "0             ABE-ORD             1\n",
      "1             ABE-PGD             2\n",
      "2             ABE-SFB             2\n",
      "3             ABI-DFW             1\n",
      "4             ABQ-ATL             1\n",
      "...               ...           ...\n",
      "3150          VPS-SHV             1\n",
      "3151          XNA-ATL             4\n",
      "3152          XNA-DEN             1\n",
      "3153          XNA-ORD             1\n",
      "3154          YAK-CDV             1\n",
      "\n",
      "[3155 rows x 2 columns]\n"
     ]
    }
   ],
   "source": [
    "# Testing the function\n",
    "\n",
    "train_flights_subset['origin_dest_pair'] = train_flights_subset['origin'] + '-' + train_flights_subset['dest']\n",
    "airport_traffic = train_flights_subset.groupby('origin_dest_pair').size().reset_index(name='flight_count')\n",
    "print(airport_traffic)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 761,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>...</th>\n",
       "      <th>dep_hour_of_day</th>\n",
       "      <th>dep_time_of_day</th>\n",
       "      <th>arr_hour_of_day</th>\n",
       "      <th>arr_time_of_day</th>\n",
       "      <th>origin_state</th>\n",
       "      <th>destination_state</th>\n",
       "      <th>arr_delay_mean_by_mkt_unique_carrier</th>\n",
       "      <th>arr_delay_mean_by_op_unique_carrier</th>\n",
       "      <th>arr_delay_encoded</th>\n",
       "      <th>origin_dest_pair</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3153</td>\n",
       "      <td>19</td>\n",
       "      <td>2100</td>\n",
       "      <td>3153</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>14</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>NY</td>\n",
       "      <td>PA</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.953181</td>\n",
       "      <td>0</td>\n",
       "      <td>SYR-PHL</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>114</td>\n",
       "      <td>4</td>\n",
       "      <td>1997</td>\n",
       "      <td>114</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>12478</td>\n",
       "      <td>...</td>\n",
       "      <td>23</td>\n",
       "      <td>evening</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>CA</td>\n",
       "      <td>NY</td>\n",
       "      <td>6.360829</td>\n",
       "      <td>6.360829</td>\n",
       "      <td>1</td>\n",
       "      <td>OAK-JFK</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>3276</td>\n",
       "      <td>18</td>\n",
       "      <td>3831</td>\n",
       "      <td>3276</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>12264</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "      <td>WV</td>\n",
       "      <td>DC</td>\n",
       "      <td>10.231880</td>\n",
       "      <td>7.331876</td>\n",
       "      <td>0</td>\n",
       "      <td>CKB-IAD</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>283</td>\n",
       "      <td>23</td>\n",
       "      <td>3806</td>\n",
       "      <td>283</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "      <td>MD</td>\n",
       "      <td>OH</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>1</td>\n",
       "      <td>BWI-CVG</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1021</td>\n",
       "      <td>23</td>\n",
       "      <td>3001</td>\n",
       "      <td>1021</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>11292</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>7</td>\n",
       "      <td>morning</td>\n",
       "      <td>IN</td>\n",
       "      <td>CO</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>2.438350</td>\n",
       "      <td>1</td>\n",
       "      <td>IND-DEN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6995</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>660</td>\n",
       "      <td>1</td>\n",
       "      <td>443</td>\n",
       "      <td>660</td>\n",
       "      <td>13303</td>\n",
       "      <td>MIA</td>\n",
       "      <td>Miami, FL</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>9</td>\n",
       "      <td>morning</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "      <td>FL</td>\n",
       "      <td>PA</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.396410</td>\n",
       "      <td>0</td>\n",
       "      <td>MIA-PHL</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6996</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1984</td>\n",
       "      <td>1</td>\n",
       "      <td>438</td>\n",
       "      <td>1984</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>Philadelphia, PA</td>\n",
       "      <td>13303</td>\n",
       "      <td>...</td>\n",
       "      <td>10</td>\n",
       "      <td>morning</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>PA</td>\n",
       "      <td>FL</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.396410</td>\n",
       "      <td>0</td>\n",
       "      <td>PHL-MIA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6997</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3337</td>\n",
       "      <td>17</td>\n",
       "      <td>2350</td>\n",
       "      <td>3337</td>\n",
       "      <td>11057</td>\n",
       "      <td>CLT</td>\n",
       "      <td>Charlotte, NC</td>\n",
       "      <td>12197</td>\n",
       "      <td>...</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "      <td>13</td>\n",
       "      <td>afternoon</td>\n",
       "      <td>NC</td>\n",
       "      <td>NY</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>6.229542</td>\n",
       "      <td>0</td>\n",
       "      <td>CLT-HPN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6998</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>445</td>\n",
       "      <td>1</td>\n",
       "      <td>414</td>\n",
       "      <td>445</td>\n",
       "      <td>14893</td>\n",
       "      <td>SMF</td>\n",
       "      <td>Sacramento, CA</td>\n",
       "      <td>11057</td>\n",
       "      <td>...</td>\n",
       "      <td>22</td>\n",
       "      <td>evening</td>\n",
       "      <td>6</td>\n",
       "      <td>morning</td>\n",
       "      <td>CA</td>\n",
       "      <td>NC</td>\n",
       "      <td>5.328596</td>\n",
       "      <td>5.396410</td>\n",
       "      <td>0</td>\n",
       "      <td>SMF-CLT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6999</th>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3050</td>\n",
       "      <td>18</td>\n",
       "      <td>2532</td>\n",
       "      <td>3050</td>\n",
       "      <td>13487</td>\n",
       "      <td>MSP</td>\n",
       "      <td>Minneapolis, MN</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>8</td>\n",
       "      <td>morning</td>\n",
       "      <td>11</td>\n",
       "      <td>morning</td>\n",
       "      <td>MN</td>\n",
       "      <td>OH</td>\n",
       "      <td>2.144038</td>\n",
       "      <td>7.331876</td>\n",
       "      <td>0</td>\n",
       "      <td>MSP-CVG</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>6496 rows × 41 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
       "0                      0            0                3153                 19   \n",
       "1                      2            2                 114                  4   \n",
       "2                      8            8                3276                 18   \n",
       "3                     10           10                 283                 23   \n",
       "4                     10           10                1021                 23   \n",
       "...                  ...          ...                 ...                ...   \n",
       "6995                   0            0                 660                  1   \n",
       "6996                   0            0                1984                  1   \n",
       "6997                   0            0                3337                 17   \n",
       "6998                   0            0                 445                  1   \n",
       "6999                   3            3                3050                 18   \n",
       "\n",
       "      tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0         2100               3153              15096    SYR   \n",
       "1         1997                114              13796    OAK   \n",
       "2         3831               3276              11027    CKB   \n",
       "3         3806                283              10821    BWI   \n",
       "4         3001               1021              12339    IND   \n",
       "...        ...                ...                ...    ...   \n",
       "6995       443                660              13303    MIA   \n",
       "6996       438               1984              14100    PHL   \n",
       "6997      2350               3337              11057    CLT   \n",
       "6998       414                445              14893    SMF   \n",
       "6999      2532               3050              13487    MSP   \n",
       "\n",
       "             origin_city_name  dest_airport_id  ... dep_hour_of_day  \\\n",
       "0                Syracuse, NY            14100  ...              13   \n",
       "1                 Oakland, CA            12478  ...              23   \n",
       "2     Clarksburg/Fairmont, WV            12264  ...               6   \n",
       "3               Baltimore, MD            11193  ...               8   \n",
       "4            Indianapolis, IN            11292  ...               6   \n",
       "...                       ...              ...  ...             ...   \n",
       "6995                Miami, FL            14100  ...               9   \n",
       "6996         Philadelphia, PA            13303  ...              10   \n",
       "6997            Charlotte, NC            12197  ...              11   \n",
       "6998           Sacramento, CA            11057  ...              22   \n",
       "6999          Minneapolis, MN            11193  ...               8   \n",
       "\n",
       "     dep_time_of_day  arr_hour_of_day  arr_time_of_day origin_state  \\\n",
       "0          afternoon               14        afternoon           NY   \n",
       "1            evening                8          morning           CA   \n",
       "2            morning                7          morning           WV   \n",
       "3            morning               10          morning           MD   \n",
       "4            morning                7          morning           IN   \n",
       "...              ...              ...              ...          ...   \n",
       "6995         morning               11          morning           FL   \n",
       "6996         morning               13        afternoon           PA   \n",
       "6997         morning               13        afternoon           NC   \n",
       "6998         evening                6          morning           CA   \n",
       "6999         morning               11          morning           MN   \n",
       "\n",
       "      destination_state  arr_delay_mean_by_mkt_unique_carrier  \\\n",
       "0                    PA                              5.328596   \n",
       "1                    NY                              6.360829   \n",
       "2                    DC                             10.231880   \n",
       "3                    OH                              2.438350   \n",
       "4                    CO                              2.438350   \n",
       "...                 ...                                   ...   \n",
       "6995                 PA                              5.328596   \n",
       "6996                 FL                              5.328596   \n",
       "6997                 NY                              5.328596   \n",
       "6998                 NC                              5.328596   \n",
       "6999                 OH                              2.144038   \n",
       "\n",
       "      arr_delay_mean_by_op_unique_carrier  arr_delay_encoded  origin_dest_pair  \n",
       "0                                5.953181                  0           SYR-PHL  \n",
       "1                                6.360829                  1           OAK-JFK  \n",
       "2                                7.331876                  0           CKB-IAD  \n",
       "3                                2.438350                  1           BWI-CVG  \n",
       "4                                2.438350                  1           IND-DEN  \n",
       "...                                   ...                ...               ...  \n",
       "6995                             5.396410                  0           MIA-PHL  \n",
       "6996                             5.396410                  0           PHL-MIA  \n",
       "6997                             6.229542                  0           CLT-HPN  \n",
       "6998                             5.396410                  0           SMF-CLT  \n",
       "6999                             7.331876                  0           MSP-CVG  \n",
       "\n",
       "[6496 rows x 41 columns]"
      ]
     },
     "execution_count": 761,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Analysing some results\n",
    "\n",
    "train_flights_subset.groupby('origin_dest_pair').tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From this result, it is noted that the origin_dest_pair of the airports OAK-JFK have the same mean of 'arrival_delay' by 'mk_unique_carrier' and 'op_unique_carrier' (both airlines that sell the ticket and operate the flight). In addition to that, the pair CKB-IAD have a high mean of arrival delay caused by the carrier that sells the ticket ('mkt_unique_carrier')."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 762,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0  1  2  3  4  5  6  7  8  9 10]\n"
     ]
    }
   ],
   "source": [
    "# Checking the labeling encoding order\n",
    "\n",
    "label_encoder = LabelEncoder()\n",
    "label_encoder.fit(train_flights_subset['mkt_unique_carrier'])\n",
    "\n",
    "# Getting the unique classes and their order\n",
    "\n",
    "classes = label_encoder.classes_\n",
    "print(classes)\n",
    "\n",
    "# Example to convert back the label enconding to the original variable\n",
    "\n",
    "# label_encoder = LabelEncoder()\n",
    "\n",
    "# train_flights_subset['mkt_unique_carrier'] = label_encoder.fit_transform(train_flights_subset['mkt_unique_carrier'])\n",
    "# train_flights_subset['op_unique_carrier'] = label_encoder.fit_transform(train_flights_subset['op_unique_carrier'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 763,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mkt_unique_carrier                        int64\n",
       "mkt_carrier                               int64\n",
       "mkt_carrier_fl_num                        int64\n",
       "op_unique_carrier                         int64\n",
       "tail_num                                  int64\n",
       "op_carrier_fl_num                         int64\n",
       "origin_airport_id                         int64\n",
       "origin                                   object\n",
       "origin_city_name                         object\n",
       "dest_airport_id                           int64\n",
       "dest                                     object\n",
       "dest_city_name                           object\n",
       "arr_delay                               float64\n",
       "diverted                                  int64\n",
       "dup                                      object\n",
       "crs_elapsed_time                          int64\n",
       "actual_elapsed_time                     float64\n",
       "air_time                                float64\n",
       "flights                                   int64\n",
       "distance                                  int64\n",
       "carrier_delay                           float64\n",
       "weather_delay                           float64\n",
       "nas_delay                               float64\n",
       "security_delay                          float64\n",
       "late_aircraft_delay                     float64\n",
       "year                                      int64\n",
       "month                                     int64\n",
       "day_of_month                              int64\n",
       "day_of_week                               int64\n",
       "quarter                                  object\n",
       "semester                                 object\n",
       "dep_hour_of_day                           int64\n",
       "dep_time_of_day                          object\n",
       "arr_hour_of_day                           int64\n",
       "arr_time_of_day                          object\n",
       "origin_state                             object\n",
       "destination_state                        object\n",
       "arr_delay_mean_by_mkt_unique_carrier    float64\n",
       "arr_delay_mean_by_op_unique_carrier     float64\n",
       "arr_delay_encoded                         int64\n",
       "origin_dest_pair                         object\n",
       "dtype: object"
      ]
     },
     "execution_count": 763,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking some columns that are still object to be converted into numerical ones\n",
    "\n",
    "train_flights_subset.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 764,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>origin_city_name</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>...</th>\n",
       "      <th>quarter_Q3</th>\n",
       "      <th>quarter_Q4</th>\n",
       "      <th>dep_time_of_day_afternoon</th>\n",
       "      <th>dep_time_of_day_evening</th>\n",
       "      <th>dep_time_of_day_morning</th>\n",
       "      <th>dep_time_of_day_night</th>\n",
       "      <th>arr_time_of_day_afternoon</th>\n",
       "      <th>arr_time_of_day_evening</th>\n",
       "      <th>arr_time_of_day_morning</th>\n",
       "      <th>arr_time_of_day_night</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3153</td>\n",
       "      <td>19</td>\n",
       "      <td>2100</td>\n",
       "      <td>3153</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>Syracuse, NY</td>\n",
       "      <td>14100</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>114</td>\n",
       "      <td>4</td>\n",
       "      <td>1997</td>\n",
       "      <td>114</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>12478</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>3276</td>\n",
       "      <td>18</td>\n",
       "      <td>3831</td>\n",
       "      <td>3276</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>Clarksburg/Fairmont, WV</td>\n",
       "      <td>12264</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>283</td>\n",
       "      <td>23</td>\n",
       "      <td>3806</td>\n",
       "      <td>283</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>Baltimore, MD</td>\n",
       "      <td>11193</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1021</td>\n",
       "      <td>23</td>\n",
       "      <td>3001</td>\n",
       "      <td>1021</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>Indianapolis, IN</td>\n",
       "      <td>11292</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
       "0                   0            0                3153                 19   \n",
       "1                   2            2                 114                  4   \n",
       "2                   8            8                3276                 18   \n",
       "3                  10           10                 283                 23   \n",
       "4                  10           10                1021                 23   \n",
       "\n",
       "   tail_num  op_carrier_fl_num  origin_airport_id origin  \\\n",
       "0      2100               3153              15096    SYR   \n",
       "1      1997                114              13796    OAK   \n",
       "2      3831               3276              11027    CKB   \n",
       "3      3806                283              10821    BWI   \n",
       "4      3001               1021              12339    IND   \n",
       "\n",
       "          origin_city_name  dest_airport_id  ... quarter_Q3 quarter_Q4  \\\n",
       "0             Syracuse, NY            14100  ...          0          1   \n",
       "1              Oakland, CA            12478  ...          0          1   \n",
       "2  Clarksburg/Fairmont, WV            12264  ...          1          0   \n",
       "3            Baltimore, MD            11193  ...          0          0   \n",
       "4         Indianapolis, IN            11292  ...          0          0   \n",
       "\n",
       "   dep_time_of_day_afternoon  dep_time_of_day_evening dep_time_of_day_morning  \\\n",
       "0                          1                        0                       0   \n",
       "1                          0                        1                       0   \n",
       "2                          0                        0                       1   \n",
       "3                          0                        0                       1   \n",
       "4                          0                        0                       1   \n",
       "\n",
       "   dep_time_of_day_night  arr_time_of_day_afternoon  arr_time_of_day_evening  \\\n",
       "0                      0                          1                        0   \n",
       "1                      0                          0                        0   \n",
       "2                      0                          0                        0   \n",
       "3                      0                          0                        0   \n",
       "4                      0                          0                        0   \n",
       "\n",
       "   arr_time_of_day_morning  arr_time_of_day_night  \n",
       "0                        0                      0  \n",
       "1                        1                      0  \n",
       "2                        1                      0  \n",
       "3                        1                      0  \n",
       "4                        1                      0  \n",
       "\n",
       "[5 rows x 50 columns]"
      ]
     },
     "execution_count": 764,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Converting some categorical features with the .getdummies function \n",
    "# Columns to be changed: 'quarter', 'semester', 'dep_time_of_day' and 'arr_time_of_day'\n",
    "\n",
    "train_flights_subset = pd.get_dummies(train_flights_subset, columns=['quarter', 'dep_time_of_day', 'arr_time_of_day'])\n",
    "train_flights_subset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 765,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Converting 'origin_state' and 'destination_state' to categories\n",
    "\n",
    "train_flights_subset['origin_state'] = train_flights_subset['origin_state'].astype('category')\n",
    "train_flights_subset['destination_state'] = train_flights_subset['destination_state'].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 766,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mkt_unique_carrier                      0\n",
       "mkt_carrier                             0\n",
       "mkt_carrier_fl_num                      0\n",
       "op_unique_carrier                       0\n",
       "tail_num                                0\n",
       "op_carrier_fl_num                       0\n",
       "origin_airport_id                       0\n",
       "origin                                  0\n",
       "origin_city_name                        0\n",
       "dest_airport_id                         0\n",
       "dest                                    0\n",
       "dest_city_name                          0\n",
       "arr_delay                               0\n",
       "diverted                                0\n",
       "dup                                     0\n",
       "crs_elapsed_time                        0\n",
       "actual_elapsed_time                     0\n",
       "air_time                                0\n",
       "flights                                 0\n",
       "distance                                0\n",
       "carrier_delay                           0\n",
       "weather_delay                           0\n",
       "nas_delay                               0\n",
       "security_delay                          0\n",
       "late_aircraft_delay                     0\n",
       "year                                    0\n",
       "month                                   0\n",
       "day_of_month                            0\n",
       "day_of_week                             0\n",
       "semester                                0\n",
       "dep_hour_of_day                         0\n",
       "arr_hour_of_day                         0\n",
       "origin_state                            0\n",
       "destination_state                       0\n",
       "arr_delay_mean_by_mkt_unique_carrier    0\n",
       "arr_delay_mean_by_op_unique_carrier     0\n",
       "arr_delay_encoded                       0\n",
       "origin_dest_pair                        0\n",
       "quarter_Q1                              0\n",
       "quarter_Q2                              0\n",
       "quarter_Q3                              0\n",
       "quarter_Q4                              0\n",
       "dep_time_of_day_afternoon               0\n",
       "dep_time_of_day_evening                 0\n",
       "dep_time_of_day_morning                 0\n",
       "dep_time_of_day_night                   0\n",
       "arr_time_of_day_afternoon               0\n",
       "arr_time_of_day_evening                 0\n",
       "arr_time_of_day_morning                 0\n",
       "arr_time_of_day_night                   0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 766,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking again the columns list and missing values\n",
    "\n",
    "train_flights_subset.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 767,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>dest</th>\n",
       "      <th>...</th>\n",
       "      <th>quarter_Q3</th>\n",
       "      <th>quarter_Q4</th>\n",
       "      <th>dep_time_of_day_afternoon</th>\n",
       "      <th>dep_time_of_day_evening</th>\n",
       "      <th>dep_time_of_day_morning</th>\n",
       "      <th>dep_time_of_day_night</th>\n",
       "      <th>arr_time_of_day_afternoon</th>\n",
       "      <th>arr_time_of_day_evening</th>\n",
       "      <th>arr_time_of_day_morning</th>\n",
       "      <th>arr_time_of_day_night</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3153</td>\n",
       "      <td>19</td>\n",
       "      <td>2100</td>\n",
       "      <td>3153</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>114</td>\n",
       "      <td>4</td>\n",
       "      <td>1997</td>\n",
       "      <td>114</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>12478</td>\n",
       "      <td>JFK</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>3276</td>\n",
       "      <td>18</td>\n",
       "      <td>3831</td>\n",
       "      <td>3276</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>12264</td>\n",
       "      <td>IAD</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>283</td>\n",
       "      <td>23</td>\n",
       "      <td>3806</td>\n",
       "      <td>283</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>11193</td>\n",
       "      <td>CVG</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1021</td>\n",
       "      <td>23</td>\n",
       "      <td>3001</td>\n",
       "      <td>1021</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>11292</td>\n",
       "      <td>DEN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6995</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>660</td>\n",
       "      <td>1</td>\n",
       "      <td>443</td>\n",
       "      <td>660</td>\n",
       "      <td>13303</td>\n",
       "      <td>MIA</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6996</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1984</td>\n",
       "      <td>1</td>\n",
       "      <td>438</td>\n",
       "      <td>1984</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>13303</td>\n",
       "      <td>MIA</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6997</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3337</td>\n",
       "      <td>17</td>\n",
       "      <td>2350</td>\n",
       "      <td>3337</td>\n",
       "      <td>11057</td>\n",
       "      <td>CLT</td>\n",
       "      <td>12197</td>\n",
       "      <td>HPN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6998</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>445</td>\n",
       "      <td>1</td>\n",
       "      <td>414</td>\n",
       "      <td>445</td>\n",
       "      <td>14893</td>\n",
       "      <td>SMF</td>\n",
       "      <td>11057</td>\n",
       "      <td>CLT</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6999</th>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3050</td>\n",
       "      <td>18</td>\n",
       "      <td>2532</td>\n",
       "      <td>3050</td>\n",
       "      <td>13487</td>\n",
       "      <td>MSP</td>\n",
       "      <td>11193</td>\n",
       "      <td>CVG</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>6975 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
       "0                      0            0                3153                 19   \n",
       "1                      2            2                 114                  4   \n",
       "2                      8            8                3276                 18   \n",
       "3                     10           10                 283                 23   \n",
       "4                     10           10                1021                 23   \n",
       "...                  ...          ...                 ...                ...   \n",
       "6995                   0            0                 660                  1   \n",
       "6996                   0            0                1984                  1   \n",
       "6997                   0            0                3337                 17   \n",
       "6998                   0            0                 445                  1   \n",
       "6999                   3            3                3050                 18   \n",
       "\n",
       "      tail_num  op_carrier_fl_num  origin_airport_id origin  dest_airport_id  \\\n",
       "0         2100               3153              15096    SYR            14100   \n",
       "1         1997                114              13796    OAK            12478   \n",
       "2         3831               3276              11027    CKB            12264   \n",
       "3         3806                283              10821    BWI            11193   \n",
       "4         3001               1021              12339    IND            11292   \n",
       "...        ...                ...                ...    ...              ...   \n",
       "6995       443                660              13303    MIA            14100   \n",
       "6996       438               1984              14100    PHL            13303   \n",
       "6997      2350               3337              11057    CLT            12197   \n",
       "6998       414                445              14893    SMF            11057   \n",
       "6999      2532               3050              13487    MSP            11193   \n",
       "\n",
       "     dest  ...  quarter_Q3  quarter_Q4  dep_time_of_day_afternoon  \\\n",
       "0     PHL  ...           0           1                          1   \n",
       "1     JFK  ...           0           1                          0   \n",
       "2     IAD  ...           1           0                          0   \n",
       "3     CVG  ...           0           0                          0   \n",
       "4     DEN  ...           0           0                          0   \n",
       "...   ...  ...         ...         ...                        ...   \n",
       "6995  PHL  ...           0           0                          0   \n",
       "6996  MIA  ...           0           0                          0   \n",
       "6997  HPN  ...           0           1                          0   \n",
       "6998  CLT  ...           0           1                          0   \n",
       "6999  CVG  ...           0           0                          0   \n",
       "\n",
       "      dep_time_of_day_evening  dep_time_of_day_morning  dep_time_of_day_night  \\\n",
       "0                           0                        0                      0   \n",
       "1                           1                        0                      0   \n",
       "2                           0                        1                      0   \n",
       "3                           0                        1                      0   \n",
       "4                           0                        1                      0   \n",
       "...                       ...                      ...                    ...   \n",
       "6995                        0                        1                      0   \n",
       "6996                        0                        1                      0   \n",
       "6997                        0                        1                      0   \n",
       "6998                        1                        0                      0   \n",
       "6999                        0                        1                      0   \n",
       "\n",
       "      arr_time_of_day_afternoon  arr_time_of_day_evening  \\\n",
       "0                             1                        0   \n",
       "1                             0                        0   \n",
       "2                             0                        0   \n",
       "3                             0                        0   \n",
       "4                             0                        0   \n",
       "...                         ...                      ...   \n",
       "6995                          0                        0   \n",
       "6996                          1                        0   \n",
       "6997                          1                        0   \n",
       "6998                          0                        0   \n",
       "6999                          0                        0   \n",
       "\n",
       "      arr_time_of_day_morning  arr_time_of_day_night  \n",
       "0                           0                      0  \n",
       "1                           1                      0  \n",
       "2                           1                      0  \n",
       "3                           1                      0  \n",
       "4                           1                      0  \n",
       "...                       ...                    ...  \n",
       "6995                        1                      0  \n",
       "6996                        0                      0  \n",
       "6997                        0                      0  \n",
       "6998                        1                      0  \n",
       "6999                        1                      0  \n",
       "\n",
       "[6975 rows x 40 columns]"
      ]
     },
     "execution_count": 767,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dropping columns of the df that is going to be converted in the .csv file to be used in the model\n",
    "\n",
    "train_flights_subset = train_flights_subset.drop(['dup', 'diverted', 'semester', 'origin_city_name', 'dest_city_name', 'carrier_delay', 'weather_delay', 'nas_delay', 'security_delay', 'late_aircraft_delay'], axis = 1)\n",
    "train_flights_subset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 768,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Converting the df to a .csv file\n",
    "\n",
    "train_flights_subset.to_csv('train_flights_subset_cleaned_version_4.csv', index = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Selection / Dimensionality Reduction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We need to apply different selection techniques to find out which one will be the best for our problems.\n",
    "\n",
    "- Original Features vs. PCA components?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After the EAD and feature engineering analysis, we noticed that it is not an easy task to know which features should be selected to try out in our first model as we are going to lose the impact of other variables and factors to be considered in the analysis. So this feature selection is one of our first tries."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 769,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing Libraries\n",
    "\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 770,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>dest</th>\n",
       "      <th>...</th>\n",
       "      <th>quarter_Q3</th>\n",
       "      <th>quarter_Q4</th>\n",
       "      <th>dep_time_of_day_afternoon</th>\n",
       "      <th>dep_time_of_day_evening</th>\n",
       "      <th>dep_time_of_day_morning</th>\n",
       "      <th>dep_time_of_day_night</th>\n",
       "      <th>arr_time_of_day_afternoon</th>\n",
       "      <th>arr_time_of_day_evening</th>\n",
       "      <th>arr_time_of_day_morning</th>\n",
       "      <th>arr_time_of_day_night</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3153</td>\n",
       "      <td>19</td>\n",
       "      <td>2100</td>\n",
       "      <td>3153</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>114</td>\n",
       "      <td>4</td>\n",
       "      <td>1997</td>\n",
       "      <td>114</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>12478</td>\n",
       "      <td>JFK</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>3276</td>\n",
       "      <td>18</td>\n",
       "      <td>3831</td>\n",
       "      <td>3276</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>12264</td>\n",
       "      <td>IAD</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>283</td>\n",
       "      <td>23</td>\n",
       "      <td>3806</td>\n",
       "      <td>283</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>11193</td>\n",
       "      <td>CVG</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1021</td>\n",
       "      <td>23</td>\n",
       "      <td>3001</td>\n",
       "      <td>1021</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>11292</td>\n",
       "      <td>DEN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
       "0                   0            0                3153                 19   \n",
       "1                   2            2                 114                  4   \n",
       "2                   8            8                3276                 18   \n",
       "3                  10           10                 283                 23   \n",
       "4                  10           10                1021                 23   \n",
       "\n",
       "   tail_num  op_carrier_fl_num  origin_airport_id origin  dest_airport_id  \\\n",
       "0      2100               3153              15096    SYR            14100   \n",
       "1      1997                114              13796    OAK            12478   \n",
       "2      3831               3276              11027    CKB            12264   \n",
       "3      3806                283              10821    BWI            11193   \n",
       "4      3001               1021              12339    IND            11292   \n",
       "\n",
       "  dest  ...  quarter_Q3  quarter_Q4  dep_time_of_day_afternoon  \\\n",
       "0  PHL  ...           0           1                          1   \n",
       "1  JFK  ...           0           1                          0   \n",
       "2  IAD  ...           1           0                          0   \n",
       "3  CVG  ...           0           0                          0   \n",
       "4  DEN  ...           0           0                          0   \n",
       "\n",
       "   dep_time_of_day_evening  dep_time_of_day_morning  dep_time_of_day_night  \\\n",
       "0                        0                        0                      0   \n",
       "1                        1                        0                      0   \n",
       "2                        0                        1                      0   \n",
       "3                        0                        1                      0   \n",
       "4                        0                        1                      0   \n",
       "\n",
       "   arr_time_of_day_afternoon  arr_time_of_day_evening  \\\n",
       "0                          1                        0   \n",
       "1                          0                        0   \n",
       "2                          0                        0   \n",
       "3                          0                        0   \n",
       "4                          0                        0   \n",
       "\n",
       "   arr_time_of_day_morning  arr_time_of_day_night  \n",
       "0                        0                      0  \n",
       "1                        1                      0  \n",
       "2                        1                      0  \n",
       "3                        1                      0  \n",
       "4                        1                      0  \n",
       "\n",
       "[5 rows x 40 columns]"
      ]
     },
     "execution_count": 770,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Loading the cleaned and processed df\n",
    "\n",
    "train_flights_subset_cleaned_version_0 = pd.read_csv('train_flights_subset_cleaned_version_4.csv', sep = ',')\n",
    "train_flights_subset_cleaned_version_0.head() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 771,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Selecting the features to be included in the analysis\n",
    "# Adding 'arr_delay' on the y-axis\n",
    "\n",
    "X = train_flights_subset_cleaned_version_0.drop('arr_delay', axis=1).select_dtypes(include=[np.number])\n",
    "y = train_flights_subset_cleaned_version_0['arr_delay']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 772,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-1.1783183 , -1.1783183 ,  1.1119689 , ..., -0.81114181,\n",
       "        -0.62538381, -0.15322409],\n",
       "       [-0.64874131, -0.64874131, -1.49699601, ..., -0.81114181,\n",
       "         1.59901806, -0.15322409],\n",
       "       [ 0.93998968,  0.93998968,  1.21756373, ..., -0.81114181,\n",
       "         1.59901806, -0.15322409],\n",
       "       ...,\n",
       "       [-1.1783183 , -1.1783183 ,  1.2699319 , ..., -0.81114181,\n",
       "        -0.62538381, -0.15322409],\n",
       "       [-1.1783183 , -1.1783183 , -1.21283432, ..., -0.81114181,\n",
       "         1.59901806, -0.15322409],\n",
       "       [-0.38395281, -0.38395281,  1.02354396, ..., -0.81114181,\n",
       "         1.59901806, -0.15322409]])"
      ]
     },
     "execution_count": 772,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Normalizing the data to have less variance\n",
    "\n",
    "scaler = StandardScaler()\n",
    "X_scaled = scaler.fit_transform(X)\n",
    "X_scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 773,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-10 {color: black;background-color: white;}#sk-container-id-10 pre{padding: 0;}#sk-container-id-10 div.sk-toggleable {background-color: white;}#sk-container-id-10 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-10 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-10 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-10 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-10 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-10 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-10 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-10 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-10 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-10 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-10 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-10 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-10 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-10 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-10 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-10 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-10 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-10 div.sk-item {position: relative;z-index: 1;}#sk-container-id-10 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-10 div.sk-item::before, #sk-container-id-10 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-10 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-10 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-10 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-10 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-10 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-10 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-10 div.sk-label-container {text-align: center;}#sk-container-id-10 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-10 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-10\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>PCA()</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-10\" type=\"checkbox\" checked><label for=\"sk-estimator-id-10\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">PCA</label><div class=\"sk-toggleable__content\"><pre>PCA()</pre></div></div></div></div></div>"
      ],
      "text/plain": [
       "PCA()"
      ]
     },
     "execution_count": 773,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Performing PCA\n",
    "\n",
    "pca = PCA()\n",
    "pca.fit(X_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 774,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1.38628005e-01 1.30525160e-01 7.88687647e-02 6.96584352e-02\n",
      " 6.46007650e-02 6.06222551e-02 4.11277806e-02 4.06024517e-02\n",
      " 3.98699689e-02 3.58210056e-02 3.38665938e-02 3.15835150e-02\n",
      " 3.08068055e-02 3.03578049e-02 2.94932647e-02 2.92796638e-02\n",
      " 2.89964577e-02 2.64410247e-02 2.15767554e-02 1.32612775e-02\n",
      " 9.98003401e-03 7.95569160e-03 1.90554759e-03 1.52493385e-03\n",
      " 1.01191439e-03 9.66163287e-04 3.97288155e-04 2.70672338e-04\n",
      " 2.98628747e-32 3.53525871e-33 3.21222692e-33 9.44434429e-34\n",
      " 5.13897318e-34 1.16508460e-34]\n"
     ]
    }
   ],
   "source": [
    "# Getting their pricipal components and their variance\n",
    "\n",
    "pcs = pca.components_\n",
    "var_ratios = pca.explained_variance_ratio_\n",
    "print(var_ratios)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 775,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjoAAAG0CAYAAAA7Go31AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAABQv0lEQVR4nO3dd3xN5x8H8M9NcrOXDEkkCFlmJELSlhaxWkWNoKWqQ1FVRas6tapUdYpWS1E/o9TeRasoaq9YsRMZZO95x/n9EU5cI8nh3pybm8/79fJqnuec3PPNdZr78ZznPEchCIIAIiIiIhNkJncBRERERIbCoENEREQmi0GHiIiITBaDDhEREZksBh0iIiIyWQw6REREZLIYdIiIiMhkMegQERGRyWLQISIiIpNlIXcBxkIQBKjVWr29nlJpDgBQqTR6e00yLTxHqDI8R6gitf38sLAwg0KhqHy/aqilRlCrtcjOLtTb67m7OwCAXl+TTAvPEaoMzxGqSG0/P5ydbcWwVxFeuiIiIiKTxaBDREREJotBh4iIiEwWgw4RERGZLAYdIiIiMlkMOkRERGSyGHSIiIjIZDHoEBERkcli0CEiIiKTxaBDREREJkv2oJOZmYmuXbvi0KFDD9xnz5496NWrF0JCQvDMM89g165d1VghERER1VSyBp1jx45h0KBBuH79+gP3iYuLw1tvvYW3334bR48exVtvvYVx48YhJSWlGislIiKimki2h3quW7cO0dHRmDhxIsaPH1/hfm3atEGXLl0AAD169MDatWvxxx9/YOzYsdVVLhER6YEgCFBrtFCptVBpBKjV2rK2puy/arVwx9dl/apb+6g1AjRaAVqtAEEQoBUEaAVA0N7+WoBWC2iFW9tvfV3WBnB7/1tt4XYbd7Xv2G7MrKzKPsJLStQyV1IxF0dr9GrnC0dbS1mOL1vQad++PXr16gULC4sKg87ly5cRGBio0+fv74/Y2Fi91qNUmotPgtUnQ7wmmRaeI1QZQ58jgiCgpFSDolI1Sko1KC7VoLhEjeJSddnXpRqU3P665HZf+b6lKk1ZOFFrxa9Vai1UKi1Uag1K1bfatwIL1T5WVhZ4o38rWY4tW9Bxd3ev0n4FBQWwsbHR6bO2tkZhYe18LD0R0YOoNVpcS87BzYxCFBarUVSiQmGx+tYfFYpK1CgsUaPoVruwpGxbUbEKWuMevKAaro6jtWzHli3oVJWNjQ2Ki4t1+oqLi2FnZ6fX46hUGmRn6y883f4XWFpant5ek0wLzxGqTGXnSH6RCleScnA5KQeXE3Nw7UYuStU1b8REoQCU5mZQWpjBwvzWHwszKM0Vd3xd1m9upoCZmQJmCkChKP/aTKGAwkwBs1t9ilt9ZopbX9/Rp1AACtzVVtzZvr2PcbOzLwsPBfnFlewpLxdHa4QEuOn9d52zsy2USvNK9zP6oBMYGIizZ8/q9F2+fBktWrSQqSIiouqnFQTczCgsCzVJObiSlIMbGYYZ2bZSmsNKaQZLpTmsLM1vtW/9sbxj260/twOK8lYgEQOLhaIsoNzVL/731j7mZrLfAFwj8R9LVWP0Qad379747bffsHXrVnTr1g07duzA4cOH8dFHH8ldGhGRwajUGpy+nI7zcZk4dTEVV5JyUFBc+aRTczMFvN3sYGejhLWlOWysLO75r42lBaytzGFjaVHWb2UOa0sLWCvNoVSawUxh7GMZRFVnlEEnNDQUU6ZMQe/eveHn54effvoJ33zzDT766CN4e3tj9uzZaNSokdxlEhHpVWp2EU5fycDpqxmIvZ6FUlXll6EcbJXw93aCv48T/L2d4OvpAKVF5cP5RLWFQjD2++eqCefoUHXjOUIqtQYXrmcj5moGTl/NREpmxb+DFAC83e3g7+0EP28nBPg4wd3ZBgqOwNRKtf13iMnM0SEiMiUpWYU4fSUDZ65lIjY+q8LJw1aW5mjq64IG7nbw93FCYy8n2Frz1zaRFPw/hojIgDRaLc7HZeHUrUtSqVlFFe7v5WqLlo1d0dLPFe1CfaC0MK+1/2In0gcGHSIiA0hOL8C+0zdw4MxN5BSUPnA/K0tzNGtYBy0bu6JFYxe4OZWvG8a5NkSPjkGHiEhPCovVOHw+BftO38DV5NwH7uftZlc2atPYBQH1nWFhzturiQyFQYeI6BFoBQHn47OwP+YGjl1Mg+o+c27MzRRo2dgVwf6uaNnIFa5O8q0SS1TbMOgQET2E1KxC7D99E/+duYGM3JL77tPAwx7tW3rhseaesLdRVnOFRAQw6BARVVlJqQZHL6RiX8wNXEjIvu8+9jZKPNbcA+1beqGBBx/YSiQ3Bh0iokoIgoB9p29g1a4ryC9S3bPdTKFAy8YuaB/shVb+bpxzQ2REGHSIiCqQlJaPJdsv4GJizj3bvFxt0T7YC08094STvZUM1RFRZRh0iIjuo6RUg43/XcOOwwnQaMsXkLdSmuPx5h5oF+yFxl6OXJWYyMgx6BAR3eXk5XQs23ERGbnFOv1hge54oUsAXBx51xRRTcGgQ0R0S2ZuMX7/+xKOX0zT6XdzssaQroFo5e8mU2VE9LAYdIio1lNrtPj7aCI27LuGEpVG7Dc3U+DpiAbo+YQvrKrw8EAiMj4MOkRUq11OzMHi7bFITCvQ6Q+s74yh3QLh7W4vU2VEpA8MOkRUK+UXqbB69xX8eypZp9/eRolBkf54ooUnJxoTmQAGHSKqdQ6fT8HSHRfvWRPnqVZeiOroz1WMiUwIgw4R1RoqtQa//30Je07qjuL4uNthaPcgBPg4y1MYERkMgw4R1QopWYX4ed0ZXE/NF/sslWbo074xurTx4WrGRCaKQYeITN6R2FT8tvU8ikvL76hq5OWAUc+1gLuzjYyVEZGhMegQkclSqbVY+c9l7DyeqNPfOcwHAzv5Q2nBURwiU8egQ0QmKS27CD+vP4O4m3lin42VOV55pinaNKkrY2VEVJ0YdIjI5By7kIaFW8+jqEQt9jXwsMcbfVrAo46tjJURUXVj0CEik6HWaLFq1xX8dTRBp79TqDee7+wPpQVXNyaqbRh0iMgkpOcU4ZcNZ3E1OVfss7I0x8tPN0FEMw8ZKyMiOTHoEFGNd/JyOhZsPoeC4vJLVT7udnijTwt4udrJWBkRyY1Bh4hqLLVGi3X/XsWfh67r9D/VyguDuwTCkg/iJKr1GHSIqEbKL1Lh5/VncD4+S+yzVJrhpe5BeKKFl4yVEZExYdAhohonOb0A0WtikJpVJPbVcyu7VOXtxktVRFSOQYeIapSYKxmYu/EMikrKVzkOb1oXrzzTFFaWvFRFRLoYdIioRhAEAX8dScAfuy5DEMr7+z3VGM8+3hAKhUK+4ojIaDHoEJHRU6m1WLLjAvbF3BD7rJTmGN6zGcKC3GWsjIiMHYMOERm13IJS/LTuNC4l5oh9ro5WeKt/MBp4OMhYGRHVBAw6RGS0ElLzEb06Bhm5xWKfv48TxvRtCUc7SxkrI6KagkGHiIzS8Ytp+HXTOZSoyicdt2/phaHdg/jUcSKqMgYdIjIqgiBgy4F4rP33qtinUAADO/mjW9v6nHRMRJIw6BCR0ShVabDoz1gcPJci9tlYmWNk7xYI9nOVsTIiqqkYdIjIKGTlleDHtTG4diNP7KvrbIOxUcGox0UAieghMegQkezibuYienUMsvNLxb6mDevgjT4tYG+jlLEyIqrpGHSISFZHY1Mxf/M5lKq1Yl+nUG+80CUAFuacdExEj4ZBh4hkIQgCNv8Xh3V7r4l9ZgoFXugSgM5hPjJWRkSmhEGHiKpdqUqD3/6MxaE7Jh3bWlngjb4t0NzXRcbKiMjUMOgQUbXKyS/B7LWncTU5V+zzqFM26djLlZOOiUi/GHSIqNrE38xD9JoYZOWViH2cdExEhsSgQ0TV4vjFNMzbdBalqvJJxx1DvTGYk46JyIAYdIjIoARBwNaD8VizR3el48FdAhHZ2psrHRORQTHoEJHBqNQaLPrzAg6cvSn22ViZ443nWqBFY650TESGx6BDRAaRU1CKH9fG4EpS+aRjd2drvB3ViisdE1G1YdAhIr1LSM1H9OpTyMgtn3QcVN8Zb/ZryUnHRFStGHSISK9OXkrH3E1nUVKqEfueauWFF7sFcdIxEVU7Bh0i0gtBELD9cAJW7boM4VafQgEM6uSPrm3rc9IxEcmCQYeIHplao8WS7RewN+aG2GdtaY5RzzVHsJ+bjJURUW3HoENEjyS/SIU5604j9nq22OfmZI2xUcHwcbeXrzAiIjDoENEjuJlZiFmrTiElq0js8/d2wph+LeFoZyljZUREZRh0iOihnIvLxJx1Z1BYohb7Hm/ugZefaQKlhbmMlRERlWPQISLJdp9MwrIdF6HRCmJf36cao+fjDTnpmIiMCoMOEVWZVivgj38u46+jCWKfpYUZhvdshjZN6spYGRHR/THoEFGVFJWoMXfjWcRcyRD7nOwtMbZ/MBp5OcpYGRHRgzHoEFGl0rOLMGtNDJLSCsS+Bh72GNs/GC6O1jJWRkRUMQYdIqrQ5aQc/LgmBrmFKrEvNMANI3o1h5UlJx0TkXFj0CGiBzp49iYWbo2FWqMV+3o81hD9OjSGGScdE1ENwKBDRPcQBAEb98dhw75rYp+5mQLDnm6C9sFeMlZGRCQNgw4R6VBrtFj0Zyz+O3NT7LO3UeLNvi0Q1KCOjJUREUnHoENEooJiFX5aq/s4B08XW4wbEIy6dWzlK4yI6CEx6BARgLI7q75fdQo3MgrFvqD6znizX0vY2yhlrIyI6OEx6BARrt3IxazVMcgtKBX7yh7n0BRKCzMZKyMiejQMOkS13ImLaZi78SxK1eV3VvV6whd9nmzExzkQUY3HoENUi/11JAErdl7C7SdW8c4qIjI1DDpEtZBWK2DFzkv4+1ii2GdjZY43+7ZEM18XGSsjItIvyRff9+/fj1GjRqFfv35IS0vDV199BbVaLfnAGRkZGD16NNq0aYOIiAhMmzbtga/zv//9D5GRkWjdujV69eqF7du3Sz4eEZUpKdXgp3WndUKOq6MVPngxjCGHiEyOpKCzadMmTJw4EUFBQYiPjwcA/PPPP/juu+8kH3jcuHGwtbXF3r17sXr1ahw4cACLFi26Z789e/Zg7ty5mD9/Po4fP44xY8Zg3LhxSExMvPdFiahCOfkl+Or34zhxKV3sa+jhgI9eagMfd3sZKyMiMgxJQWfevHmYM2cOxo8fDzMzM7i7u2Pu3LnYvHmzpIPGx8fj8OHDmDhxImxsbFC/fn2MHj0ay5Ytu2ffq1evQhAE8Y+5uTmUSiUsLHjVjUiKpPQCfLH4GOJu5ol9rfxcMWlIKJztrWSsjIjIcCSlhZs3b6JVq1YAIN6N0bBhQxQWFlb0bfe4dOkSnJ2d4eHhIfb5+fkhOTkZubm5cHR0FPufffZZrF27Fj169IC5uTkUCgW+/vpreHp6SjpmZZRKc7i7O+j1NQEY5DXJtFTHOXLqUhpmLD2GguLyy8M92zXC8D4tYW7GO6uMHX+PUEV4flRM0oiOr68vdu7cqdP333//oWHDhpIOWlBQABsbG52+2+27Q5NKpUKTJk2watUqnDx5Ep9//jk++ugjXLhwQdIxiWqrXccS8NmvB8SQo1AAw59rgRF9GXKIyPRJGtEZP348Ro8ejc6dO6OkpASfffYZNm/ejG+//VbSQW1tbVFUVKTTd7ttZ2en0z916lS0bt0awcHBAID+/ftj8+bNWLduHd5//31Jx62ISqVBdra0kamK3E7YaWl5lexJtZWhzxFBELDlQDzW/ntV7LO0MMPrvZojLMgd6en5Bjku6Q9/j1BFavv54exsC6XSvNL9JI3oPPHEE1ixYgUcHR0REREBrVaLhQsXokOHDpKKCwgIQHZ2NtLTyydEXrlyBZ6ennBw0B2CS05ORmlpqU6fhYUFlEouSU/0IBqtFou3X9AJOQ62SkwcHIqwIHcZKyMiql6Sby/PyMjA2LFjMW/ePHTu3Bn5+dL/Vejr64uwsDBMnz4d+fn5SEhIwJw5cxAVFXXPvpGRkVi6dCnOnj0LrVaLbdu24dChQ+jRo4fk4xLVBsWlasxecxp7TiaLfR51bPDR0DD41XOSsTIiouonKegsW7YMb7/9NtLS0gCUhZ5x48Zh/fr1kg8cHR0NtVqNzp07Y+DAgXjyyScxevRoAEBoaCg2btwIABgzZgyGDBmCt956C23btsW8efPw008/oWnTppKPSWTqcgpK8dXvJxBzJUPs8/N2xIdDw/j0cSKqlRSCIAiV71amS5cumDVrFpo3by72nTlzBu+8806NX8SPc3Souun7HLmRUYDvV55Cek6x2Nc60B0jejWDZRWuY5Px4e8RqkhtPz+qOkdH0mTkjIyMe0ZSmjVrhoyMjAd8BxFVh0uJ2YheHaNz+3iXMB883zkAZryziohqMUmXrvz9/bFhwwadvk2bNqFx48Z6LYqIqu5obCq+Xn5SJ+QMivTHC10YcoiIJI3ojBs3Dm+88QZWrlyJevXq4caNGzh37hzmzZtnqPqIqAI7Dl/HH/9cFp8+bmGuwPCezRDe1KPC7yMiqi0kjei0a9cOGzZsQLt27WBvb4+nnnoKmzZtQnh4uKHqI6L70GoF/P73Ray4I+TYWVvg3edDGXKIiO4g+YFRjRo1wpgxYwxRCxFVQalKg183ncOxi2lin6ujNcYPbIV6bnYVfCcRUe0jKehcunQJM2fORFxcHLRarc62ux8NQUT6l1dYitlrTuNyUo7Y19DDAeMGBMOJD+YkIrqHpKAzefJk2NjYYMSIEXx6OFE1S88uwncrT+FmZvkyCC0bu+KNPs1hbcn/H4mI7kfSb8cLFy7g33//hb29vaHqIaL7SEjNx3crTyInv/xxKE+18sLQ7kEwN5O8wDkRUa0hKejUrVv3nudOEZFhxcZnYfbaGBSVaMS+Pk82Qq8nfKFQ8PZxIqKKSAo6L774It5880289NJLcHNz09nWtm1bvRZGRGVr5MzbdBZqTdm9VQoFMOzpJniqVT2ZKyMiqhkkBZ0vvvgCAHDixAmdfoVCgfPnz+uvKiLCP8cTsWzHRfH2caWFGUY91xyhAXz6OBFRVUkKOrGxsYaqg4huEQQB6/Zexeb/4sU+O2sLjI0KRoCPs3yFERHVQJJv1SgqKkJOTo54e7lKpcLFixfRtWtXvRdHVNtotFos3nYBe2NuiH11HKwwYVAIvLlGDhGRZJKCzpo1azB16lSUlJTo9Lu6ujLoED2iEpUGczecxcnL6WKft5sdxg9sBRdHaxkrIyKquSQFnV9++QXjxo2DnZ0djhw5gmHDhuHrr79Gu3btDFUfUa2QX6RC9OoYnYUAA3ycMDYqGHbWShkrIyKq2SQtwJGWloZhw4bh8ccfx/Xr19G8eXNMnz4dq1atMlR9RCYvM7cYXy49phNyQgPc8M6gEIYcIqJHJGlEx9XVFSqVCl5eXrh27RoAoF69esjIyDBIcUSmLv5mLqYtOYasvPLLwR1C6uHFboFcCJCISA8k/SYNDg7G5MmTUVxcDF9fXyxfvhzr1q2Ds7OzgcojMl1nr2Zg0o/7dEJO73a+eImrHRMR6Y2kEZ0PPvgAH3/8MQoKCjBx4kSMGjUKxcXF+PLLLw1VH5FJOnk5Hb+sP4NSddndiwoF8GK3IHQK9Za5MiIi06IQBEGofLf7U6vVUKlUsLGx0WdNslCpNMjOLqx8xypyd3cAAKSl5entNck0HI1NxdyNZ6HRlv2vZ2FuhpG9myMsiAsBki7+HqGK1Pbzw9nZFkqleaX7VWlEZ/PmzejZsyfWr1//wH369OlT1dqIaq2DZ29i/ubz0N7694WttQXe6tcSQQ3qyFwZEZFpqlLQ+eWXX9CzZ09ER0ffd7tCoWDQIarE3lPJWPRnrPhIBwdbJT4f8QScrCv/FwkRET2cKo/oAMDMmTMRGhoKc3P+YiaS4p/jiVi646LYdrBVYtob7dConlOtHXYmIqoOkm7tePPNN1FaWmqoWohM0vbD13VCjpO9JSYNbo1G9ZxkrIqIqHaQFHTq16+P06dPG6oWIpOz6b84/PHPZbHt4miF94e0Rj0+t4qIqFpIur3cyckJr7zyCnx8fFC3bl0oFApx2+LFi/VeHFFNVfYE8mvY/F+c2OfubI2Jz4fCzbnm36VIRFRTSAo6oaGhCA0NNVQtRCZBEASs2nUF2w5fF/s8XGwx8fkQPpyTiKiaSQo6Y8aMMVQdRCZBKwhY/tcl7DyeKPZ5u9nh3edD4GRvJWNlRES1k6Sgk5WVhSVLliAlJQVabdmKriqVChcvXsTGjRsNUiBRTaHVCli8PRb/nroh9jWoa493ng+Bg62ljJUREdVekh8BERcXBxcXF+Tn56NevXrYt28fhgwZYqj6iGoEjVaLhVvO48DZFLGvkZcjJgxqxSeQExHJSFLQOXLkCLZu3YqUlBTMmzcPP/74IzZs2CCus0NUG6k1WszbdA5HY1PFPn8fJ4wf0Ao2VpL+FyMiIj2TdHu5hYUFPDw84OvriwsXLgAAnn32WZw7d84gxREZO5VaiznrzuiEnCYNnDFhIEMOEZExkBR0vL29cebMGTg6OqKgoACZmZkoLCxEcXGxoeojMlpqjRa/bDiDk5fTxb4WjV0wbkArWFsy5BARGQNJv40HDx6MoUOHYsuWLejZsyeGDRsGCwsLtG3b1lD1ERkljVaLeRvP4sSl8pAT4u+GN/q0gNJC0r8fiIjIgKoUdL788ksMHToUUVFRCAwMhJubGyZOnIjffvsNBQUFePXVVw1dJ5HR0GoFzN98HkcvpIl9If5uGN23BSzMGXKIiIxJlYLOqVOnsHTpUnTo0AEvvfQSLC3LbpUdMWKEQYsjMjZarYCFW8/j0Lnyu6taNnbFG30YcoiIjFGVfjOvWLECa9asgZubG0aPHo1evXph1apVKCkpMXR9REZDKwj437ZY/HfmptjX3LcOxvTj5SoiImOlEARBkPINBQUFWL9+PVasWIHU1FQMHDgQQ4YMgaenp6FqrBYqlQbZ2YV6ez13dwcAQFpant5ek+QjCAKW7LiI3SeSxL4mDZzx9oBWsFKaP9Rr8hyhyvAcoYrU9vPD2dkWyir8/pX8z1A7OzsMGTIEmzZtws8//4zExER06dLloYokqgkEQcDvf1/SCTkBPk4YGxX80CGHiIiqx0PfA7t//36sWrUKe/bsQceOHfVYEpHxEAQBf/xzGTuPlT+7ys/bkbeQExHVEJJ+U6empmLNmjVYvXo1CgoK0L9/f2zZsgX16tUzVH1EshEEAWv2XMWOIwliXyMvB4wfEMLFAImIaogq/bbetWsXVq5cib1798LPzw+jRo1C7969YWXFpzGT6dqw7xq2HowX2w087DFhUAhsrRlyiIhqiir9xh4zZgw6d+6M3377jYsDUq2waf81bNwfJ7Z93O3w7vOhfEAnEVENU6Wgs3Pnzhp/VxVRVf15MB7r9l4T2/XcykKOvQ1DDhFRTVOlu64Ycqi22HEkAat2XxHbni62mPh8CBztLGWsioiIHhZXOSO6ZdfxRKzYeUls161jg4kvhMLJnnPRiIhqKgYdIgBHY1OxdMdFse3mZI33XghFHQeGHCKimoxBh2q9y4k5+HXzOdxeItzF0QrvvRAKF0drWesiIqJHV6XJyEOHDoVCoahwn8WLF+ulIKLqdDOzENFrYqBSawEAtlYWmDAwBG7ONjJXRkRE+lClEZ2IiAiEh4ejXr16OHfuHJo2bYru3bujVatWuHDhAho1amToOon0LregFN+vPIn8IhUAwMJcgbf6t0Q9NzuZKyMiIn2p8jo6ADB48GDMmzcPrVu3Frd1794dn3zyiWGqIzKQklINZq2OQVp2sdj32rPNENSgjoxVERGRvkmao3P+/Hm0atVKpy8oKAhxcXH6rInIoLRaAXM3nsW1G7li34BOfoho5iFjVUREZAiSgo6fnx8WLVqk0/fLL7+gSZMm+qyJyGAEQcCyvy/i5OV0sS+ytTeeDm8gY1VERGQokh7a8+GHH2LUqFFYsmQJPD09kZycDK1WiwULFhiqPiK92nboOnYdTxLboQFuGNwlsNLJ9kREVDNJCjqtW7fGjh07sHv3bqSkpMDT0xORkZFwcHAwVH1EenPw3E2dVY8b13PEiN7NYWbGkENEZKokP4bZ2dkZgYGBsLW1RceOHZGXl2eIuoj06sL1LCzccl5s13W2wdioYFgpzWWsioiIDE3SHJ2MjAw8//zzGDhwICZNmoSEhAR06dIFJ06cMFR9RI8sKb0As9echlpTtiSgvY0S4we2gqMtn19FRGTqJAWd6dOnIzAwEEeOHIGFhQX8/PwwYsQIzJw501D1ET2SrLwS/LDyJApL1AAApYUZxkYFw8PFVubKiIioOkgKOgcPHsQHH3wAGxsbcfLm8OHDcfnyZYMUR/QoikrUmLXqFDJySwAACgAjejWHv7eTvIUREVG1kRR0lEoliovLFlgThLLLAAUFBbCz40qyZFzUGi1+Xn8G11Pzxb7nuwQgLMhdxqqIiKi6SQo6kZGRmDhxIuLi4qBQKJCRkYEpU6agQ4cOhqqPSDJBELB4+wWcuZYp9nVrWx9d29SXsSoiIpKDpKDzzjvvwNbWFk8//TRyc3PRvn17FBUV4d133zVUfUSSbfovDvtibojtNk3qYmCkv4wVERGRXCTdXm5nZ4fo6GhkZmYiMTERnp6eqFu3rqFqI5LswNmbWL/3mtgO8HHC6z2bwowLAhIR1UqS19FJTU3F9evXIQgC4uPjER8fDwBo27at3osjkuLC9Sz8trV8rRwPF1u81T8YSguulUNEVFtJCjpLlizBjBkzoNFodPoVCgXOnz//gO8iMrwbGQX4ce1da+UMCIa9jVLmyoiISE6Sgs7//vc/TJ48Gf3794eFheTBICKDyC0sxQ+rTqGguGytHAtzM4ztH4y6dbhWDhFRbScprWRmZmLAgAEwM5M0h5nIYFRqDX5ccxpp2cVi3/CeTeHvw7VyiIhI4l1X4eHhOHTokKFqIZJEKwiYv/k8LifliH39OzRGeFMPGasiIiJjImlEx8PDAyNHjkRERATc3Nx0tn355ZeSDpyRkYFPPvkEhw8fhrm5OXr37o1Jkybd95LY4cOH8fXXX+Py5ctwdHTE4MGDMXLkSEnHI9Ozds9VHIlNFdtPtfJCj8caylgREREZG0kjOqWlpXj22WfvCTkPY9y4cbC1tcXevXuxevVqHDhwAIsWLbpnvytXrmDEiBEYPHgwjh8/jrlz52LhwoXYtm3bI9dANdeek0nYejBebDdv5IIXuwWJjyYhIiICAIVw+1kO1Sg+Ph7dunXDv//+Cw+PsssMW7duxddff41du3bp7Dt16lRkZ2fj22+/FfuuXbsGe3t7uLvrbzl/lUqD7OxCvb2eu7sDACAtLU9vr0llzlzLwA8rY6C9dep6u9vhgyFhsLWuWRPkeY5QZXiOUEVq+/nh7GwLpbLy5UOq9Mkwb948jBgxAj/++OMD9xkzZkyVi7t06RKcnZ3FkAMAfn5+SE5ORm5uLhwdHcX+mJgYPPHEE5gwYQL2798PFxcXvPzyyxg0aFCVj1cVSqW5eNLokyFeszaLv5GLn9efFUNOHQcrfD7yiRp9hxXPEaoMzxGqCM+PilUp6Bw5cgQjRox44ERkqZcLCgoKYGNjo9N3u11YWKgTdHJycrB48WJ8//33mDlzJk6cOIGRI0fCyckJTz/9tKTjUs2WmVuMz+YfRFFJ2W3kVpbmmPzaYzU65BARkWFVKej8+uuvAMoWDNQHW1tbFBUV6fTdbt/9JHRLS0t07twZHTt2BFC2AvNzzz2HP//8U69Bh5eujFtJqQYzfj+O9Oyy80QBYGSv5nCyNq+x7zHPEaoMzxGqSG0/P/R66epOBw8eREpKCm5P7VGpVLhw4QI+/vjjKr9GQEAAsrOzkZ6eLk5svnLlCjw9PeHgoDsE5+fnh9LSUp0+jUYDGaYWkUy0WgFzN55F/M3y/5mf7xKAkIBHnxRPRESmTVLQ+eKLL7BixQpx1EWj0aCgoABPPvmkpIP6+voiLCwM06dPx+eff46srCzMmTMHUVFR9+z7/PPPY/jw4diwYQN69+6No0ePYtOmTfjmm28kHZNqrhU7L+Hk5XSx3SXMB13b1JexIiIiqikk3V7+559/YunSpfjhhx8QGRmJI0eOYNiwYfD09JR84OjoaKjVanTu3BkDBw7Ek08+idGjRwMAQkNDsXHjRgDA448/jjlz5mDx4sUICwvDBx98gEmTJqFz586Sj0k1z19HE/D3sUSxHeLvhuc7B8hYERER1SSSRnSKiooQEhKCtLQ0nD17FgqFAmPGjEGPHj0kH9jNzQ3R0dH33XbixAmddocOHdChQwfJx6Ca7eSldKzYeUlsN/R0wMjezWFmxrVyiIioaiSN6Hh6eiIjIwPu7u64efMmVCoVrK2tkZ+fb6j6qJa6npKHuRvP4vZULFdHK7wdFQwry8onnhEREd0maUSnQ4cOePnll/G///0Pbdu2xYcffggrKyv4+voaqDyqjbLySjBrdQxKVBoAgLWlOd6OagVneyuZKyMioppG0ojOhAkT8Nxzz0GpVGLy5MnIysrC5cuXMXXqVEPVR7VMSakG0atjkJVXAgBQKIA3+rSAT117mSsjIqKaSJZHQBgjrqMjP61WwE/rTuPEpfI7rF7sFojI1j4yVmU4PEeoMjxHqCK1/fzQ6zo6H3zwQaX7SH16OdHdVu++ohNyurTxMdmQQ0RE1UPSpSsiQ9l9MgnbDl8X28F+rng+kreRExHRo6nSiA5Ha8iQzsZlYun2i2K7fl173kZORER6IWlER61WIzo6Gp07d0ZwcDC6d++OBQsWGKo2qgWS0gswZ90Z8WnkTvaWeDsqGDZWkp9OQkREdA9JnyYzZ87E7t27MXLkSHh5eSEhIQELFy5ESUmJuKoxUVXlFpRi1qpT4tPILZVmeDsqGC6O1jJXRkREpkJS0Nm4cSNWrlyJBg0aiH2PPfYYhg0bxqBDkqjUGsxeG4P0nGIAZU8jH9GrOXw9HeUtjIiITIqkS1eCIMDd3V2nz8fHh08SJ0kEQcCCLedxJSlX7BvQyR+tA90r+C4iIiLpJAWdIUOG4OOPP0ZubtkHVElJCWbMmHHfp44TPciGfddw+Hyq2O4QUg/dw/k0ciIi0j9Jl67WrFmDlJQU/Pnnn3ByckJeXh7U6rL5FXPnzhX3O3/+vH6rJJNx4MxNbNwfJ7ab+dbBkK6BUCh4hxUREemfpKDz9ddfG6oOqgUuJmTjtz/LQ7CXqy1G92kBC3Mu50RERIYhKeg4OTkhKCjonv7NmzejZ8+eeiuKTE9KViF+XHsaak3ZfC4HWyXGDWgFW2ulzJUREZEpk/RP6UGDBmHVqlViu7CwEJMmTcLHH3+s98LIdKg1Wvy09jTyi1QAAAtzM7zVLxjuzjYyV0ZERKZO0ojODz/8gA8++ABHjhxBVFQUPv74Yzg5OWHt2rWGqo9MwO4TSUhMKxDbrz7bBP4+TjJWREREtYWkEZ2OHTtiw4YNOHr0KIYNG4awsDD88ccfaNy4saHqoxour7AU6/deE9tPtfLCY808ZayIiIhqE0lBJzU1FZ988gmKi4vx/PPP488//8T8+fO5jg490Pp911B4a+VjGytz9HvKT+aKiIioNpEUdHr27Am1Wo0NGzbg008/xcKFC/HHH3/ghRdeMFR9VIMlpuZj94kksd27XSM42lnKWBEREdU2koLO8OHDsWDBAnF15NatW2P9+vXw9vY2SHFUcwmCgN//vojbg30eLrboHOYjb1FERFTrSAo6I0aMuKfPwcEB77//vt4KItNw/GI6Yq9ni+3nI/25Xg4REVW7Kn3y9OrVS6f966+/6rS7d++uv4qoxlOpNfjjn0tiu0VjFwT7ucpYERER1VZVCjqJiYk67buDDicj0512HEkQn0pubqbA85EBfMQDERHJokpB5+4PqbuDDT/E6LasvBJs/i9ebEe29kE9NzsZKyIiotrsoSZNMNjQg6zZcwUlKg0AwN5Gid7tfeUtiIiIajXODiW9uZKcg//O3BTb/Z5qDDs+y4qIiGTEoEN6oRUELP+7fAKyj7s9nmpVT8aKiIiIqvisq8LCQjRt2lRsC4Kg0yY6ePYmribniu3BXQJgZsZLnEREJK8qBZ3Fixcbug6qwYpL1Vi1+4rYbhPkjiYN68hYERERUZkqBZ3w8HBD10E12JYD8cjJLwUAWJibYWAnf5krIiIiKsM5OvRIUrOLsP1wgth+OqIB3JxtZKyIiIioHIMOPZJV/1yGWqMFANRxsMKzjzWUuSIiIqJyDDr00M7HZeLYxTSxHdXRD1aW5jJWREREpOuhg05mZqY+66AaRqPVYvnO8tvJ/bwd8VgzDxkrIiIiupekoKNWq/H9998jLCwMkZGRSEhIQP/+/ZGammqo+shI/XvqBhLTCsT24C6BXDGbiIiMjqSgM3v2bBw8eBCzZs2CUqmEq6srPD09MW3aNEPVR0aooFiFdf9eFdvtWniikZejjBURERHdX5VuL79t06ZNWL58OTw8PKBQKGBra4svv/wSXbt2NVR9ZIQ27LuG/CIVAMDK0hz9O/rJXBEREdH9SRrRKSwshIuLC4DyJ5hbW1vDzIxzmmuLpPQC/HMsSWz3fLwhnO2tZKyIiIjowSQllJCQEPz4448Ayp9gvmTJErRs2VL/lZFRWrXrMrS3Qq67szW6ta0vc0VEREQPJunS1YcffoiXX34Z69atQ0FBAXr06IGCggL89ttvhqqPjEhsfBZirmSI7YGdAqC04O3kRERkvCQFnQYNGmDLli3YtWsXkpOT4enpiY4dO8Le3t5Q9ZGREARB53lW/t5OaB3oJmNFRERElZMUdKZOnYoBAwagR48ehqqHjNTRC2m4dqP86eQDO/nzdnIiIjJ6kuboZGRkYNCgQejXrx+WL1+OvLw8Q9VFRkSt0WLNnvLRnNAAN/j7OMlYERERUdVICjo//PAD9u3bh6ioKKxbtw5PPvkkJk2ahCNHjhiqPjICe04mIzWrCABgplAgireTExFRDSH5vnAHBwcMHjwYK1euxM8//4wjR47gpZdeMkRtZASKStTYtP+a2H6ylRe8XO1krIiIiKjqJM3RAYCCggJs27YN69evR0xMDDp27IipU6caojYyAtsPX0duYdnigJYWZujdrpHMFREREVWdpKDzzjvv4J9//oGnpycGDBiAWbNmiQsIkunJyS/B9sMJYrtbeH3UceDigEREVHNICjoWFhb49ddf0aZNG0PVQ0Zk4/44lKg0AAB7GyWeiWgoc0VERETSSAo6X331laHqICNzM7MQe04mi+1e7XxhYyX5SicREZGsqvTJ1bp1axw/fhxNmjR54Nop58+f12thJK+1e66Ij3pwc7JGp1BvmSsiIiKSrkpBZ968eQCAxYsXG7QYMg5XknNw9EKa2O7XoTEszPngViIiqnmqFHRuz8nZsWMHPv7443u2v/feewgPD9dvZSQLQRCwalf54oANPR0Q3tRDxoqIiIgeXqVBJyUlBQcOHAAArFq1Ci1atNDZnpeXh7/++ssw1VG1O3UlAxcTssX2gI5+MOOjHoiIqIaqNOjUqVMHS5cuRWZmJkpLSxEdHa2z3crKCmPGjDFYgVR9tFoBa+54cGeLRi5o5svlA4iIqOaqNOhYWlpi9erVAIDXXnsNCxYsMHhRJI/9Z24gKb0AAKAA+KgHIiKq8STNML1fyFGr1Th37pzeCiJ5lKo0WL+3/FEPjzX3QAMPBxkrIiIienSSFkbZs2cPPvvsM6SkpEC4desxULaQ4OnTp/VeHFWfv48lIiuvBABgYa5A3ycby1wRERHRo5MUdL7++mt069YNjo6OuHDhAnr27ImffvoJUVFRhqqPqkF+kQpbDsSL7cjWPnBztpGxIiIiIv2QdOkqISEBEydOxLPPPousrCx069YN3377LVauXGmo+qgabP4vDkUlagCAjZUFej7hK29BREREeiIp6Li4uMDMzAz16tXDlStld+f4+/vj5s2bBimODC89pwj/HE8U2z0eawB7G6WMFREREemPpKATFBSEWbNmAQBcXV2xZ88eHDp0CFZWfKJ1TbXu32tQa8rmW9VxsELXNvVlroiIiEh/JAWdiRMn4u+//0ZaWhrGjh2L0aNH4+WXX8Zrr71mqPrIgK6n5OHg2fLRuD7tG8FSaS5jRURERPolaTKyn58ftmzZAgDw9vbGrl27UFBQgEaNGhmkODKs1buv4Pa9c95udmjX0kvWeoiIiPStSkHnyJEjFW5PT09H27Zt9VIQVY/Y+CycuZYptvt39IOZGR/1QEREpqVKQWfo0KEVblcoFDh//rxeCiLDEwQB6/eVLw4Y6OOEVn6uMlZERERkGFUKOrGxsYaug6pRbHyWzoM7+z7VGAo+uJOIiEyQpDk6ycnJD9xWr169Ry6GDO/u0ZwmDZwR1KCOjBUREREZjqSgExkZCYVCIT7+4c5RAKmXrjIyMvDJJ5/g8OHDMDc3R+/evTFp0iRYWDy4pIsXL2LAgAGYN28eIiIiJB2PypyPz8KlxByx3YePeiAiIhMmKejs3LlTp52ZmYn58+ejc+fOkg88btw4eHh4YO/evUhPT8cbb7yBRYsWYfjw4ffdv6ioCO+88w6Ki4slH4vK3D2a07RhHQTWd5avICIiIgOTtI6Ot7e3zp+WLVviiy++QHR0tKSDxsfH4/Dhw5g4cSJsbGxQv359jB49GsuWLXvg90yZMgVdunSRdBzSdS4uC5fvGM15rj2XBSAiItMmaUTnQXJzcyXtf+nSJTg7O8PDw0Ps8/PzQ3JyMnJzc+Ho6Kiz//r16xEfH49p06Zhzpw5+ij5HkqlOdzdHfT+uoZ4zYchCAK2LD8htkMC3NGuNVdBNgbGco6Q8eI5QhXh+VExSUHnxx9/1GmrVCrs3bsXISEhkg5aUFAAGxvdp2PfbhcWFuoEnStXruD777/H8uXLYW7OVXsf1omLaYiNzxLbg7s3kbEaIiKi6iEp6Bw6dEinbW5ujtDQUIwcOVLSQW1tbVFUVKTTd7ttZ2cn9pWUlGD8+PH48MMPDX5Xl0qlQXZ2od5e73bCTkvL09trPixBELB481mx3byRC9zslUZRW21mTOcIGSeeI1SR2n5+ODvbQlmFxxZJCjpLlix56ILuFBAQgOzsbKSnp8PNzQ1A2ciNp6cnHBzKh+BOnz6NuLg4fPTRR/joo4/E/lGjRuG5557DZ599ppd6TN2Za5m4klx+eZFzc4iIqLaQFHRUKhW2bt2KpKQkaLVanW1jxoyp8uv4+voiLCwM06dPx+eff46srCzMmTMHUVFROvu1adMGMTExOn1BQUH45ZdfeHt5FQmCgPV7y++0atHIBf7eTjJWREREVH0kBZ133nkHhw4dQkBAgM4aOg+zqm50dDQ+//xzdO7cGWZmZujTpw9Gjx4NAAgNDcWUKVPQu3dvya9Luk5fzcS1G3eM5jzJ0RwiIqo9JAWdffv2YePGjfDx8XnkA7u5uT3wtvQTJ07ctx8ALly48MjHri0EQcCGfVfFdsvGrvCrx9EcIiKqPSSto+Pu7g5nZ2cDlUL6FnMlA9dulE9S49wcIiKqbSSN6EyaNAlvv/02Bg8efM9aN23bttVrYfRoykZzyufmBPu5onE9xwq+g4iIyPRICjqnTp3C/v37sX//fp1+hUIh+VlXZFinrmQg7iZHc4iIqHaTFHR+//13zJs3D+3bt4eZmaSrXlSN7h7NaeXnikZeHM0hIqLaR1JasbKyQrt27RhyjNzJy+mIv2M0pzdHc4iIqJaSlFheeeUVfPPNN8jJyal8Z5LF3aM5If5uHM0hIqJaS9Klq2XLliE5ORmLFi26Zxvn6BiHk5fScT0lX2xzbg4REdVmkoLOjBkzDFUH6cHdozmhAW5o6Mmn2hIRUe0lKeiEh4cbqg7Sg+MX03E9tXw0p3c7juYQEVHtJinoREZGPvBxDzt37tRLQfRwtIKAjfs5mkNERHQnSUHnrbfe0mlnZmZizZo1GDBggF6LIulOXExDQirn5hAREd1JUtDp27fvPX1du3bFhAkT8Morr+itKJJGe9fcnLBAdzTw4GgOERHRIy+I4+3tjbi4OD2UQg/r+IU0JKYViG2um0NERFRG0ojOkSNHdNoqlQrbtm2Dr6+vPmsiCYS75uaEBbmjfl17GSsiIiIyHpKCztChQ3XaZmZm8PPzw6effqrXoqjqzsZl6ozmPMc7rYiIiESSgk5sbKyh6qCHtONwgvh1sJ8rfDiaQ0REJKryHB1BEHD9+nWdvq1bt0Kj0ei9KKqapLR8nLmWKba7t60vYzVERETGp0pBp7CwEC+88AJmzpwp9mVkZOD999/H0KFDUVhYaLAC6cH+Olo+mlO/rj2aNKwjYzVERETGp0pB5+eff4ZSqcSUKVPEPldXV+zatQtqtRpz5841WIF0f7kFpfjvTIrY7ta2/gMXcyQiIqqtqhR0tm/fji+++AKurq46/a6urpgyZQq2bdtmkOLowXadSIJaowUAONlZIryph8wVERERGZ8qBZ2MjAw0bNjwvtuaNm2KtLQ0vRZFFVOpNdh1PFFsR4b5QGnxyEsiERERmZwqfTra29sjKyvrvtuys7NhY2Oj16KoYgfPpiC3UAUAUFqYoWNIPZkrIiIiMk5VCjqPP/44li1bdt9tv//+O0JCQvRZE1VAEATsOFI+CbldC0842FrKWBEREZHxqtI6OiNHjkS/fv2QlZWFHj16wN3dHampqfjzzz+xZs0aLF261NB10i1n4zKRlF6+QGBX3lJORET0QFUKOo0aNcKCBQvw6aefYtmyZVAoFBAEAYGBgfj111/RokULQ9dJt9w5mhPs5wovVzsZqyEiIjJuVV4ZuXXr1ti0aRMSEhKQmZkJd3d31KvHuSHVKSktH2euli8Q2I2jOURERBWS9AgIAKhfvz7q1+cHrBzuXCDQx90eTblAIBERUYV4T3INkVuou0Bg93AuEEhERFQZBp0aYvfx8gUCHblAIBERUZUw6NQAKrUG/9yxQGDn1t5cIJCIiKgK+GlZAxw8d9cCgaHeMldERERUMzDoGLm7Fwh8ggsEEhERVRmDjpE7F5eFpLQ7FghswzveiIiIqopBx8htP3Jd/DrYzxX13LhAIBERUVUx6BixpPQCnQUC+bgHIiIiaRh0jNhfR+5cINAOzbhAIBERkSQMOkaqbIHAm2K7W9sGXCCQiIhIIgYdI7X7hO4CgRHNuEAgERGRVAw6Rkil1uCfY+ULBEZygUAiIqKHwk9PI8QFAomIiPSDQcfICIKgMwn5iRaecOQCgURERA+FQcfInIvPQiIXCCQiItILBh0js+Nw+WhOy8ZcIJCIiOhRMOgYkeT0Apy+miG2u4VzNIeIiOhRMOgYkV0nksSvuUAgERHRo2PQMSKnr5SP5nRq7cMFAomIiB4Rg46RSMsuQmp2kdhu5ecqYzVERESmgUHHSJyLK394p6eLLVwcrWWshoiIyDQw6BiJs3FZ4tfNfDk3h4iISB8YdIyAVhBw/o4Rnea+LjJWQ0REZDoYdIzA9ZQ8FBSrAQBmCgWCGnBEh4iISB8YdIzA2WvlozmN6jnA1tpCxmqIiIhMB4OOETh3x/wcXrYiIiLSHwYdmZWqNLiUmCO2mzHoEBER6Q2DjswuJmZDrdECAKwszdG4nqPMFREREZkOBh2ZnbtWftmqSX1nWJjzr4SIiEhf+KkqszsXCmzWiJetiIiI9IlBR0a5BaW4npovtjkRmYiISL8YdGR0Lr58NMfZ3hJerrYyVkNERGR6GHRkdPdt5XxaORERkX4x6MhEEATOzyEiIjIwBh2Z3MwsRGZuidjm+jlERET6x6AjkzsvW/m428HJzlLGaoiIiEwTg45MdC5bcTSHiIjIIBh0ZKDRahF7/Y6JyJyfQ0REZBAMOjK4lpyHohINAMDCXIFAH2d5CyIiIjJRDDoyuPOylb+3E6wszWWshoiIyHQx6MjgLOfnEBERVQvZgk5GRgZGjx6NNm3aICIiAtOmTYNarb7vvsuXL0f37t0RGhqK7t27Y9myZdVcrf4UlahxNTlXbDPoEBERGY5sQWfcuHGwtbXF3r17sXr1ahw4cACLFi26Z7+///4b3333Hb766iscP34cM2bMwA8//IDt27dXf9F6cCEhGxqtAACws7aAr6eDzBURERGZLlmCTnx8PA4fPoyJEyfCxsYG9evXx+jRo+87UpOSkoLXX38dISEhUCgUCA0NRUREBI4cOSJD5Y/u3LXyy1ZNGtaBmRkf+0BERGQoFnIc9NKlS3B2doaHh4fY5+fnh+TkZOTm5sLR0VHsHzJkiM73ZmRk4MiRI/jggw/0WpNSaQ53d/2Prtz9mhcSs8WvI1rWM8gxqWbhOUCV4TlCFeH5UTFZRnQKCgpgY2Oj03e7XVhY+MDvS0tLw+uvv44WLVqgZ8+eBq3RENKzi5CQki+2QwLcZayGiIjI9MkyomNra4uioiKdvtttOzu7+37PyZMn8fbbb6NNmzb48ssvYWGh39JVKg2ysx8csqS6nbDT0vLEvv2nb4hfuzlZw0LQ6myn2uV+5wjRnXiOUEVq+/nh7GwLpbLy5VlkGdEJCAhAdnY20tPTxb4rV67A09MTDg73DsGtXr0aL7/8MoYNG4Zvv/0WlpY187lQd95WztWQiYiIDE+WoOPr64uwsDBMnz4d+fn5SEhIwJw5cxAVFXXPvtu3b8dnn32G2bNn49VXX5WhWv0QBEHnQZ68rZyIiMjwZLu9PDo6Gmq1Gp07d8bAgQPx5JNPYvTo0QCA0NBQbNy4EQDw448/QqPRYOzYsQgNDRX/TJ48Wa7SH0pSWgFyC0oBAAoATRvWkbcgIiKiWkCWOToA4Obmhujo6PtuO3HihPj1pk2bqqskg7rzslVDTwfY2yhlrIaIiKh24CMgqgkf+0BERFT9GHSqgUqtxcWEbLHd3JeXrYiIiKoDg041uJKUg1KVFgBgaWEGfx9neQsiIiKqJRh0qsGdl60C6ztDacG3nYiIqDrwE7canOP8HCIiIlkw6BhYQbEKcTfKV61sxvk5RERE1YZBx8DOx2VBuPW1o60SPnXtZa2HiIioNmHQMbA7L1s19XWBmUIhYzVERES1C4OOgek+9oGXrYiIiKoTg44B3cwoQGp2+VPam3MiMhERUbVi0DGgkxfTxK+9XG3h4mgtYzVERES1D4OOAd0ZdJo15GgOERFRdWPQMRCNVkDM5TuCTiPOzyEiIqpuDDoGcjUpG3mFKgCAmUKBJg0YdIiIiKobg46B3HnZqnE9R9hYWchYDRERUe3EoGMgOvNzeFs5ERGRLBh0DKBEpcG5a+ULBTZvxInIREREcmDQMYDrKXlQa7QAAGtLczTycpS5IiIiotqJQccAnO2tYGlR9tZGNPOAhTnfZiIiIjlwhqwBuDvb4Ju3n0Jiaj4a17WTuxwiIqJai0HHQBrVc0Kjek5IS8uTuxQiIqJai9dUiIiIyGQx6BAREZHJYtAhIiIik8WgQ0RERCaLQYeIiIhMFoMOERERmSwGHSIiIjJZDDpERERkshh0iIiIyGQx6BAREZHJYtAhIiIik8WgQ0RERCaLQYeIiIhMlkIQBEHuIoyBIAhQq7V6ez2l0hwAoFJp9PaaZFp4jlBleI5QRWr7+WFhYQaFQlHpfgw6REREZLJ46YqIiIhMFoMOERERmSwGHSIiIjJZDDpERERkshh0iIiIyGQx6BAREZHJYtAhIiIik8WgQ0RERCaLQYeIiIhMFoMOERERmSwGHSIiIjJZDDpERERkshh0iIiIyGQx6BAREZHJYtAhIiIik8WgQ0RERCaLQYeIiIhMFoOOAWRkZGD06NFo06YNIiIiMG3aNKjVarnLIiOQmZmJrl274tChQ2LfqVOnMGDAAISGhiIyMhKrVq2SsUKSQ2xsLF555RWEh4ejXbt2eO+995CZmQmA5weVOXDgAAYMGIDWrVujXbt2mDp1KoqLiwHwHKmUQHr34osvCu+8845QWFgoXL9+XXj22WeFX3/9Ve6ySGZHjx4VunTpIgQGBgoHDx4UBEEQsrOzhfDwcGHp0qWCSqUS/vvvPyE0NFQ4deqUzNVSdSkqKhLatWsnzJo1SygpKREyMzOF119/XRg5ciTPDxIEQRAyMjKEli1bCmvWrBE0Go2QkpIi9OzZU5g1axbPkSrgiI6excfH4/Dhw5g4cSJsbGxQv359jB49GsuWLZO7NJLRunXr8O6772L8+PE6/Tt27ICzszOGDBkCCwsLPP744+jVqxfPl1okOTkZTZo0wZtvvglLS0vUqVMHgwYNwpEjR3h+EADAxcUF//33H/r16weFQoHs7GyUlJTAxcWF50gVMOjo2aVLl+Ds7AwPDw+xz8/PD8nJycjNzZWxMpJT+/bt8ddff6FHjx46/ZcuXUJgYKBOn7+/P2JjY6uzPJJR48aNMX/+fJibm4t927dvR/PmzXl+kMje3h4A0KFDB/Tq1Qvu7u7o168fz5EqYNDRs4KCAtjY2Oj03W4XFhbKURIZAXd3d1hYWNzTf7/zxdramudKLSUIAr7//nvs2rULH330Ec8PuseOHTvw77//wszMDGPHjuU5UgUMOnpma2uLoqIinb7bbTs7OzlKIiNmY2MjTii8rbi4mOdKLZSfn4+xY8di06ZNWLp0KYKCgnh+0D2sra3h4eGBiRMnYu/evTxHqoBBR88CAgKQnZ2N9PR0se/KlSvw9PSEg4ODjJWRMQoMDMSlS5d0+i5fvoyAgACZKiI5XL9+Hf3790d+fj5Wr16NoKAgADw/qMzx48fx9NNPo7S0VOwrLS2FUqmEv78/z5FKMOjoma+vL8LCwjB9+nTk5+cjISEBc+bMQVRUlNylkRHq2rUr0tPTsWjRIqhUKhw8eBCbNm1C//795S6NqklOTg6GDRuG1q1bY8GCBXBxcRG38fwgAAgKCkJxcTG+/fZblJaWIikpCV999RWioqLQvXt3niOVUAiCIMhdhKlJT0/H559/jkOHDsHMzAx9+vTBu+++qzPZkGqvoKAgLF68GBEREQCA06dPY9q0abh48SJcXFwwevRo9OvXT+Yqqbr89ttvmDFjBmxsbKBQKHS2nThxgucHASgbpZk+fTpOnz4NBwcH9OrVS7xTj+dIxRh0iIiIyGTx0hURERGZLAYdIiIiMlkMOkRERGSyGHSIiIjIZDHoEBERkcli0CEiIiKTxaBDREREJotBh4iI9Eaj0SAhIUHuMohEDDpEehIUFIQRI0bg7jU4165di8jISIMcMzIyEmvXrjXIa1fF1q1b8fjjjyMsLAy7du26Z3tQUBCCg4MRGhqK0NBQhISEIDIyEvPnz3/gax49ehShoaGPXNsvv/yC4cOHP/LrAGU/x6FDhx64PTU1FVOmTEFkZCRCQ0PRsWNHfPrppzrPvKstxo8fj/Xr18tdBpGIQYdIj/bs2VPhh7ipWbVqFZ599lkcO3YMnTp1uu8+v/76K06cOIETJ07g5MmTmDFjBmbNmoU1a9bcd/82bdrgxIkTj1zbqFGjquXvIj4+Hr169YJWq8WyZctw4sQJrFixAtnZ2XjhhReQn59v8BqMSVZWltwlEOlg0CHSo6FDh2LWrFk4fvz4fbcnJiYiKCgIiYmJYt/s2bMxdOhQAGWjP4MHD8ZXX32F8PBwPPbYY1iyZAlWrlyJTp06ISwsDJMnT9Z5zbNnz6Jfv34IDw/Ha6+9hri4OHHb9evXMWrUKERERKBTp074/vvvxScgr127Fv369cOrr76KNm3aYNOmTffUm5WVhU8++QTt27dHREQERo4cKb5+VFQUDh48iBUrVqBLly5Vfo/Cw8MREBCAc+fOASgblZo8eTLatWuHPn364MCBA+LTu2+/X6tWrUJkZCTCwsLwyiuv4ObNm+Lrbdq0CT179kRoaCieeeYZbN269b7v68CBAzF58mS0bt0a7du3x5w5c8TRt5SUFIwbNw6RkZFo1aoVOnfujNWrV1fp55k2bRqCg4MxZcoUeHl5AQA8PT0xc+ZMNG3aVHyydEXv5e2fc/369ejUqRNCQkLwwQcf4OjRo+jduzdCQ0MxbNgwZGZmAgDef/99fPjhh3jppZcQEhKCZ555Bn///bdYU1JSEsaNG4fHH38c7dq1wzvvvIPU1FQAwKFDhxAZGYmff/4ZTz75JMLDw/HWW2/pBLItW7agV69eCAsLQ79+/bBv3z5x29ChQ/Htt99iyJAh97znH330EY4ePYq5c+di1KhR4t9Dhw4dEB4ejv79+2Pnzp1Vel+J9EYgIr0IDAwUDh48KHz++edChw4dhKysLEEQBGHNmjVCp06dBEEQhISEBCEwMFBISEgQvy86Olp48cUXxX0DAwOF3377TdBoNMKyZcuEpk2bChMmTBAKCwuFmJgYoWnTpsLhw4cFQRCETp06CU899ZQQGxsrFBcXC5MnTxa6desmqFQqoaCgQOjUqZPwzTffCMXFxUJycrIQFRUlfPPNNzrHWrt2rVBSUiIUFRXd8zO9+OKLwksvvSSkpqYKRUVFwowZM4QOHToIeXl54vbo6OhK35PbSkpKhG3btgnNmzcXdu7cKf4Mzz33nJCTkyPk5OQIBw8eFAIDA3Xer9GjRws5OTlCWlqa0LNnT+GTTz4RBEEQDh48KLRo0ULYvXu3oNFohD179gjNmzcXLl26dN/3dfr06UJxcbEQExMjhIeHCytXrhQEQRCGDx8uvPvuu0JhYaGgVquFhQsXCsHBwUJ+fv59f447f54mTZoI69ate+B7UJX38vbPOW7cOKGwsFC4cOGC0LRpU6F3797CzZs3hYyMDKFr167C7NmzBUEQhEmTJglNmjQRtmzZIqhUKmHdunVC8+bNhcuXLwulpaVCt27dhAkTJgi5ublCTk6OMGHCBKFv376CSqUS399PP/1UKCoqEuLi4oR27doJc+fOFQRBEHbv3i2EhYUJhw8fFtRqtfDPP/8IISEhwsWLF8WfIzw8XDh79qxQUlIifPfdd0JYWJhQXFx8zzlx4MABoV27dkJKSoqg1WqF5cuXCxEREUJpaWml7xeRvnBEh0jPJk2aBBcXF7z//vv3zNepCltbWwwbNgxmZmZo3749NBoNXnvtNdjY2KBly5aoW7cukpKSxP1fffVVBAUFwcrKCu+//z4SExMRExOD3bt3o7S0FBMmTICVlRW8vLzw9ttvY9myZeL3KpVKPPfcc7C0tIS1tbVOHQkJCTh8+DA++eQTuLu7w9raGu+++y7UajX27NlT5Z9n1KhRaNOmDdq0aYMnnngCc+fOxSeffKIzb6l79+5wdHSEo6PjfV/j9ddfh6OjI9zc3BAZGSmOhKxfvx7dunVDhw4dYGZmhqeeegq///47PDw87nkNZ2dnvPvuu7CyskLLli0xaNAgbNy4EQDwxRdf4NNPP4VSqURycjLs7OxQXFyMnJycCn+2nJwcaLVauLu7V7hfVd/LV199FTY2NggMDIS7uzv69u0LDw8PuLi4ICQkROfvvWPHjujRowcsLCzQp08ftGjRAlu3bsXRo0eRkJCAKVOmwMHBAY6OjpgyZQpiY2Nx5swZ8fvffPNNWFtbo2HDhoiIiMC1a9cAAEuXLsULL7yAtm3bwtzcHJ06dUJkZCRWrFih8/fVrFkzWFpaom/fvsjLy0NGRsY9P7eVlRVycnKwcuVKnDt3DgMGDMCBAwegVCorfL+I9MlC7gKITI2lpSV++OEH9O3bFwsXLkSdOnUkfb+zszMUCgUAwMys7N8idwYAMzMzaLVase3j4yN+bWNjA2dnZ6SkpCApKQmZmZlo27atuF0QBKhUKvFDyd3dXTzG3W5PpK1fv77YZ25uDi8vL50P3Mr88ssviIiIqHCfunXrVrjdzc1N/NrCwkIMkKmpqWjWrJnOvsHBwfd9DW9vb50PWC8vL2zfvh1AWRCZOXMm4uLi4Ovri4YNGwKAzvt8P87OzlAqlUhLS7vv9oyMDLi4uFT6XrZq1Up8vTu33/33fmdw9vX11TmWl5cX0tLSkJGRgTp16sDe3l7cZm9vD2dnZyQlJYnv5Z3hTKlUiq+dlJSEw4cPY/ny5eJ2jUaDxx57TGzf+b0WFmUfI/d7r0JDQzF79mwsWbIE8+fPh7W1NYYOHYo33njjgecdkb4x6BAZQIMGDTB16lS899576Nevn9hvbm4OAFCpVGLf3ZM3b4ecqro99wIA8vPzkZWVBW9vb6jVajRo0ADbtm3T2X77w7eyY3l7ewMom+cTEBAAoOwDLzk5udIRDKmk/sy3eXl5ITk5Wadv4cKFCAkJuWff1NRUCIIgHisxMRH16tWDSqXCyJEjMWHCBAwePBgKhQJnzpwRR3sqolQq0b59e2zduhV9+vTR2VZaWornnnsOQ4cORd++fQFU/l5KeR9SUlJ02omJiYiMjIS3tzeysrKQn58vhp28vDxkZWXB3d290lFGT09P9OnTByNGjBD7kpOT7xnxq4rk5GS4urpiwYIFKC0txYEDBzBmzBg0b94cHTt2lPx6RA+DkZrIQHr06IH+/fvjjz/+EPtcXV3h5OSELVu2QBAEnD17VieIPIyFCxfi6tWrKCoqwrRp09C0aVO0aNECnTp1QkFBAebPn4/S0lLk5uZi0qRJGD9+fJU+UOvWrYsOHTrgiy++QFpaGoqLi/HNN99Ao9E88A6r6ta3b1/89ddf2LdvH7RaLfbu3YvZs2fDwcHhnn3T0tIwb948qFQqxMTEYNWqVRgwYABUKhWKi4thbW0NhUKB5ORkfP311wB0A+mDvPfeezh27BimTp0qho+4uDiMGTMGdnZ2GDRokEHey7/++gv//fcf1Go1Vq9ejYsXL6Jnz55o2bIl/P398emnnyIvLw95eXn47LPP0KBBA7Ru3brS1x04cCAWL16MmJgYAMDp06fRr18/bN68uUp1WVpaIi8vT/ze4cOHIzY2FpaWlnB1dQUAyaOcRI+CIzpEBvThhx/i1KlTyM3NBVD2ITB16lRER0djwYIFaNGiBQYOHIhjx4499DG6dOmCUaNGISsrC23btsWcOXNgZmYGe3t7LFq0CDNmzMD8+fOh1WoRERGBn3/+ucqvPXPmTHzzzTfo27cvCgsLERISgv/97386l1jkFBYWhq+++gpfffUVkpKS4O3tje+++w4BAQH3BEh3d3ckJiaiffv2sLOzw9tvv40ePXoAAKZPn45Zs2bhiy++gKurKwYOHIjLly/j4sWLaNSoUYU1NG7cGKtXr8ZPP/2EqKgo5Ofnw9nZGR06dMD06dPF96qi9/JhbkFv06YNfv31V4wZMwa+vr6YN2+eeGls7ty5mDFjBrp3747S0lI88cQT+O2338TLTBV5+umnUVhYiA8//BDJyclwdnbGyy+/LN7BVpk+ffrgs88+w5kzZ/D7778jLi4Ob7zxBrKysuDq6ooPP/xQvFRHVB0UwsPMliQiqkHWrl2LH3/8Ef/884/cpejF+++/DwCYMWOGzJUQGT9euiIiIiKTxaBDREREJouXroiIiMhkcUSHiIiITBaDDhEREZksBh0iIiIyWQw6REREZLIYdIiIiMhkMegQERGRyWLQISIiIpPFoENEREQm6/+2UhfwFhY7pAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the PCA to ge better understanding of the components \n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "plt.plot(np.cumsum(var_ratios))\n",
    "plt.xlabel('Number of Principal Components', fontsize=10)\n",
    "plt.xticks(fontsize=10)\n",
    "plt.ylabel('Cumulative Explained Variance', fontsize=10)\n",
    "plt.yticks(fontsize=10)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 776,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "19"
      ]
     },
     "execution_count": 776,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Extracting the components that retain the most variance\n",
    "# The treshold set is of to maintain 95% of variance\n",
    "\n",
    "n_components = np.argmax(np.cumsum(var_ratios) >= 0.95) + 1\n",
    "n_components"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 777,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-2.06131590e+00, -8.87976973e-01, -8.72534969e-01, ...,\n",
       "        -1.37045693e+00,  9.08162072e-01,  1.03844440e-01],\n",
       "       [ 5.41707281e+00,  2.07786583e+00, -1.18716831e-03, ...,\n",
       "         1.28790940e+00,  8.72156390e-01, -1.08121343e-01],\n",
       "       [-3.92265628e+00,  1.41186611e+00,  1.51403538e+00, ...,\n",
       "        -9.91213492e-01,  1.05199937e+00, -3.64506204e-01],\n",
       "       ...,\n",
       "       [-1.90547391e+00,  5.85756762e-01, -8.28857453e-01, ...,\n",
       "         5.91679801e-02, -5.86487728e-01,  1.24986720e+00],\n",
       "       [ 3.88249641e+00,  1.82680229e+00, -1.08601892e+00, ...,\n",
       "         1.22096645e+00, -5.89621909e-01, -3.50360719e-01],\n",
       "       [-2.23553092e+00,  1.67535931e+00, -4.89047969e-01, ...,\n",
       "        -2.37852668e-01, -1.84931308e-01, -5.89521136e-01]])"
      ]
     },
     "execution_count": 777,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Transforming the data\n",
    "\n",
    "X_pca = pca.transform(X_scaled)[:, :n_components]\n",
    "X_pca"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have reduced even more our features to be used (40 to 19), we will move forward to our modelling task."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modelling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use different ML techniques to predict each problem.\n",
    "\n",
    "- linear / logistic / multinomial logistic regression\n",
    "- Random Forest\n",
    "- Naive Bayes\n",
    "- SVM\n",
    "- XGBoost\n",
    "- The ensemble of your own choice"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For each model, we are going to use 'X_pca' to compare both of their functionality and results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 778,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing libraries\n",
    "\n",
    "import pickle\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "import statsmodels.api as sm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear Regression Model and Ridge Regression Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 779,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Linear Regression Model Results are:\n",
      "Mean Absolute Error (MAE) is 18.87.\n",
      "Mean Squared Error (MSE) is 3218.57.\n",
      "Root Mean Squared Error (RMSE) is 56.73.\n",
      "R2_score is 0.16.\n"
     ]
    }
   ],
   "source": [
    "# Modelling the Linear Regression Model first\n",
    "# Just to remember, y = train_flights_subset['arr_delay']\n",
    "\n",
    "X_pca_train, X_pca_test, y_train, y_test = train_test_split(X_pca, y, test_size=0.2, random_state=100)\n",
    "\n",
    "lr = LinearRegression()\n",
    "lr.fit(X_pca_train, y_train)\n",
    "y_pred_lm = lr.predict(X_pca_test)\n",
    "\n",
    "# Calculating some evaluation metrics \n",
    "\n",
    "mae = mean_absolute_error(y_test, y_pred_lm)\n",
    "mse = mean_squared_error(y_test, y_pred_lm)\n",
    "r2 = r2_score(y_test, y_pred_lm)\n",
    "rmse = np.sqrt(mse)\n",
    "\n",
    "print('The Linear Regression Model Results are:')\n",
    "print('Mean Absolute Error (MAE) is {:.2f}.'.format(mae))\n",
    "print('Mean Squared Error (MSE) is {:.2f}.'.format(mse))\n",
    "print('Root Mean Squared Error (RMSE) is {:.2f}.'.format(rmse))\n",
    "print('R2_score is {:.2f}.'.format(r2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 780,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1395.000000\n",
       "mean        4.975020\n",
       "std        24.674106\n",
       "min       -24.153602\n",
       "25%       -15.069776\n",
       "50%       -10.332886\n",
       "75%        34.351252\n",
       "max        54.303328\n",
       "dtype: float64"
      ]
     },
     "execution_count": 780,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Statistics of the y_pred_lm variable\n",
    "\n",
    "pd.Series(y_pred_lm).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 781,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkoAAAHMCAYAAAAqIofBAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAACMv0lEQVR4nOzdd3iT1fvH8XeSjrS0pWUWkT0dQCutIlBAAZUtqF+V4QARcKCgFAUV/CkiBUVQmSLgAkQQFBREnAwRkKECMmRDmR10JE2T/P6oDQ1taAqFlvJ5XReX9sl5npz0NE/unHOfcwxOp9OJiIiIiORiLOoKiIiIiBRXCpREREREPFCgJCIiIuKBAiURERERDxQoiYiIiHigQElERETEAwVKIiIiIh4oUBIRERHxQIGSiEgJpfWERS6eAiURuaR69epFvXr13P7Vr1+fxo0bc99997F06dJCf86FCxdSr149Dh065LHMoUOHqFevHgsXLiz053/hhRe4/fbb833uc38nkZGRdOvWjS+++OKi6zl58mRmzJhxwa9BRLL4FHUFRKTku/766xkxYoTrZ7vdTnx8PLNmzWLw4MEEBwfTokWLQnu+Vq1aMW/ePCpUqFBo17wUBgwYQKtWrYCs3p/U1FTmz5/P8OHDyczM5IEHHrjga7/zzjs89dRThVRTkauXAiURueSCgoKIiIjIdbxly5bceuutLFiwoFADpTJlylCmTJlCu96lUrVq1Vy/l6ZNm7Jjxw5mzZp1UYGSiBQODb2JSJHx8/PD19c31/H58+fToUMHbrzxRlq1asW7775LZmam6/HTp0/z/PPP06xZMxo0aECXLl1YtGiR6/G8ht6+++47OnfuTMOGDenatSs7duxwe85169ZRr1491q1b53a8V69e9OrVy/WzxWLhrbfe4o477uDGG2/kpptu4tFHH2X79u0X++sAwGg0ct1113HkyBGPZfbt28fAgQNp1qwZERER9OrVi40bN7oer1evHgDvvfee6/9F5MIoUBKRS87pdJKZmen6Z7Va2b9/Py+99BKpqal06dLFVXbq1Km8/PLL3HrrrUyZMoUePXowffp0XnnlFVeZIUOGsHv3bl599VWmTZvG9ddfz9ChQ3MFOdl++OEHBg4cSJ06dXjvvfdo164dQ4YMuaDXEhsbyxdffMHjjz/Ohx9+yAsvvMDOnTsZNGhQoSVP7927l6pVq+b52O7du+nWrRsHDx7kpZdeYty4cRgMBh5++GF+//13AObNmwfAvffe6/p/EbkwGnoTkUtu/fr13HDDDW7HDAYDdevWZcKECa7E5zNnzjB58mTuv/9+XnrpJQCaN29OaGgoL730Eo8++ih16tTh999/54knnqBNmzYA3HLLLYSGhmIymfJ8/vfff58bbriBt956C8A1zJf9s7cyMjJITU3l5Zdfpn379gDcfPPNpKam8uabb3LixIkC5UU5HA5XT5nD4eDYsWN8/PHH7Nixwy2nK6f33nsPX19fPvroI4KDg4GsnKyOHTsyduxY5s+f7xrOCw8Pz3PIU0S8p0BJRC65G264gVdffRWAY8eOMWHCBGw2G+PHj6dWrVqucps2bSI9PZ3bb7/dbagtO5BavXo1derU4ZZbbuHdd99lx44dtGzZkhYtWjB06NA8n9tisfD3338zcOBAt+Pt2rUrcKDk5+fnmkl2/Phx9u/fz7///suPP/4IgM1mK9D1hg8fzvDhw92OBQUF0b9/f4/5Sb///ju33XabK0gC8PHxoUOHDrz//vukpqZSqlSpAtVDRDxToCQil1ypUqVo0KABAA0aNCAyMpIuXbrQu3dvvvzyS1fidWJiIgCPP/54ntc5fvw4AOPHj2fKlCl8++23LFu2DKPRSNOmTRk5ciRVqlRxOycpKQmn05kruftCZ8T9+uuvvPHGG/z777+UKlWKevXquQKTgg69PfXUU65Zb0ajkeDgYCpXruyxZwyyXk+5cuVyHS9XrhxOp5OUlBQFSiKFSIGSiFx2ZcuW5ZVXXuHpp59m1KhRrp6dkJAQAMaNG0f16tVznZcdIAQHBzNkyBCGDBnCv//+y8qVK5k0aRKvvvoqH3zwgds5oaGhGI1GTp486XY8OyjLZjAYgKwhsJxy9tAcOHCAJ598ktatWzN16lRXHtGnn37Kr7/+WuDfQ+XKlV0BpLdKly6d67UAnDhxAoCwsLAC10NEPFMyt4gUiTvuuIOYmBiWLFniSsJu1KgRvr6+HDt2jAYNGrj++fr68tZbb3Ho0CEOHz5My5YtWbZsGQA1a9akb9++NG3alPj4+FzP4+/vT2RkJN99951bj88PP/zgVi4oKAiAo0ePuo4lJSWxZ88e189//fUXVquVfv36uSVbZwdJl2Ml7OjoaH788UfOnDnjOma321m6dCkNGjTAz88PyOqhEpGLpx4lESkyw4YNo3Pnzrz++ut8+eWXhIWF8dhjjzFhwgRSUlK45ZZbXDlNBoOB+vXrExwcTHh4OK+//jopKSlUrVqVv/76i59//pl+/frl+TyDBw/m4Ycf5qmnnuL+++9n3759TJ482a1MvXr1qFSpEu+99x7BwcEYjUamTZtGQECAq8wNN9yAj48PY8eOpXfv3mRkZLBw4UJ++uknANLS0i7Z7yrbU089xS+//MJDDz3E448/jp+fH5988gkHDx50600LCQlh06ZNrF+/nqioKFePmYgUjL5yiEiRqVmzJr169WLnzp188sknADz77LO88MILrFixgr59+zJ27FgaN27MJ5984kpgfu+994iJiWHChAn07t2bOXPm8NRTT/Hkk0/m+TxRUVFMnz6dY8eO8dRTTzF37lzeeOMNtzImk4mJEydSoUIFBg8ezOuvv067du244447XGWqVavGW2+9xbFjxxgwYIBryYKPP/4Yg8HAhg0bLsWvyU2dOnX47LPPKFeuHMOGDWPIkCE4nU4++ugjmjZt6irXv39//vzzT/r27evWSyYiBWNwatdEERERkTypR0lERETEAwVKIiIiIh4oUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIiIi4oECJREREREPtDJ3IXE6nWRmOvIvWEz4+mZtummz2Yu4JnKx1JYlh9qy5FBbFn8+PkavVqxXoFRIMjMdJCZe+u0LCkv58lkrHF9JdZa8qS1LDrVlyaG2LP5CQwNdAe35aOhNRERExAMFSiIiIiIeKFASERER8UCBkoiIiIgHCpREREREPFCgJCIiIuKBAiURERERD7SOkoiIXLVMJiNGowGHw4ndfuUsGiyXj3qURETkqmIwgJ/ZhDnEQJopgQOWPaSZEjAHG/Azm/BisWa5iqhHSURErhomkwHfQFiyazFxa+LYHL/Z9VhEeASxTWPpWKcTtjSw251FV1EpNtSjJCIiVwWDAXwDIfb7IXRf2N0tSALYHL+Z7gu7M3RlLL6BqGdJAAVKIiJylfD1N7Fk19dM2TjlvOUmb5jM0l1L8PXPfx8wKfkUKImIyFXB6OdgzOoxXpWNWxOH0U/J3aJASURErgImk5EkSyJbjm3xqvym+E0kWRIxmfQxebXTX4CIiJR4RqOB0+kJBTonwZKI0ahEpaudAiURESnxHA4nZQLCCnROmDkUh0Mz3652CpRERKTEs9sdlDaHEhEe4VX5yPBISptDtQilKFASEZGrgyPDSGzTWK/KxjaNxZGhj0hRoCQiIlcJm9VOxzqd6N+4/3nLDYgaQIc6HbFZ7ZepZlKcKVASEZGrgtMJtjSIazOWOd3m5BqGiwiPYE63OYxpHYctLau8iLYwERGRq4bd7sSRAu2qd6ZD3Y4kWRJJsCQSZg6ltDkUR4YRa4pdQZK4KFASEZGritMJGRY7WCDQFEaQfxkcdieWZAeg4TZxp0BJRESuWna7A7tiIzkP5SiJiIiIeKBASURERMQDBUoiIiIiHihQEhEREfFAgZKIiIiIBwqURERERDxQoCQiIiLigQIlEREREQ8UKImIiIh4oEBJRERExAMFSiIiIiIeKFASERER8UCBkoiIiIgHCpREREREPFCgJCIiIuKBAiURERERDxQoiYiIiHigQElERETEAwVKIiIiIh4oUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIiIi4oECJREREREPFCiJiIiIeKBASURERMSDYhMoTZo0iV69erkd2759Oz179iQiIoJWrVoxY8YMt8cdDgcTJ04kJiaGRo0a0bt3b/bv31+ga4iIiIh4UiwCpVmzZjFx4kS3YwkJCTz66KNUr16dBQsW8PTTTzNhwgQWLFjgKjNp0iTmzp3L66+/zrx58zAYDPTt25eMjAyvryEiIiLiiU9RPvmxY8cYPnw4GzdupEaNGm6Pff755/j5+TFy5Eh8fHyoVasW+/fvZ/r06dxzzz1kZGTw4YcfMmTIEFq2bAnA+PHjiYmJYcWKFXTo0CHfa4iIiIicT5EGSn///TelS5fmq6++4v333+fw4cOuxzZs2EB0dDQ+Pmer2KRJE6ZOncqpU6c4fPgwqampNGnSxPV4SEgI119/PevXr6dDhw75XqNs2bKF9lp8fU2ULx9caNe7XK7EOkve1JYlh9qy5FBbXvmKNFC6/fbbuf322/N8LD4+nrp167odq1ChAgBHjhwhPj4egEqVKuUqc/ToUa+uUZiBkoiIiJQ8RRoonY/FYsHPz8/tmL+/PwBWq5X09HSAPMskJSV5dY3CZLPZSUxMK9RrXkrZ33JOnDhTxDWRi6W2LDnUliWH2rL4Cw0NxNfXlG+5YpHMnRez2exKys6WHdwEBgZiNpsB8iwTEBDg1TVEREREzqfYBkrh4eEcP37c7Vj2zxUrVnQNueVVJjw83KtriIiIiJxPsQ2UoqOj2bhxI3a73XVs7dq11KhRg7Jly1K/fn2CgoJYt26d6/Hk5GS2bdtGVFSUV9cQEREROZ9iGyjdc889pKSkMHz4cHbv3s3ChQuZPXs2/fr1A7Jyk3r27Mm4ceNYuXIlO3bsYNCgQYSHh9O2bVuvriEiIiJyPsU2mbts2bJ88MEHjBo1iq5du1K+fHliY2Pp2rWrq8zAgQPJzMzkpZdewmKxEB0dzYwZM1wJ3N5cQ0RERMQTg9PpdBZ1JUoCzXqToqK2LDnUliWH2rL4u+JnvYmIiIgUNQVKIiIiIh4oUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIiIi4oECJREREREPFCiJiIiIeKBASURERMQDBUoiIiIiHhTbvd5ERERKGpPJiNFowOFwYrc7iro64gX1KImIiFxCBgP4mU2YQwykmRI4YNlDmikBc7ABP7MJg6Goayjnox4lERGRS8RkMuAbCEt2LSZuTRyb4ze7HosIjyC2aSwd63TClgZ2u/aoL47UoyQiInKJ+AZC7PdD6L6wu1uQBLA5fjPdF3Zn6MpYfANRz1IxpUBJRETkEsh0ZLJk19dM2TjlvOUmb5jM0l1L8PU3XaaaSUEoUBIREbkErJlWxqwe41XZuDVxGP2U3F0cKVASERG5BBItiWw5tsWrspviN5FkScRk0sdycaMWERERuQQSLAkFLJ+I0ahEpeJGgZKIiMglEGYOK2D5UBwOzXwrbhQoiYiIXAKh5lAiwiO8KhsZHklpc6gWoSyGFCiJiIhcAv4+/sQ2jfWqbGzTWBwZ+kgujtQqIiIil4CP0YeOdTrRv3H/85YbEDWADnU6YrPaL1PNpCAUKImIiFwitjSIazOWOd3m5BqGiwiPYE63OYxpHYctDZxKTyqWtIWJiIjIJWK3O3GkQLvqnelQtyNJlkQSLImEmUMpbQ7FkWHEmmJXkFSMKVASERG5hJxOyLDYwQKBpjCC/MvgsDuxJDsADbcVdwqURERELhO73YFdsdEVRTlKIiIiIh4oUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIiIi4oECJREREREPFCiJiIiIeKBASURERMQDBUoiIiIiHihQEhEREfFAgZKIiIiIBwqURERERDxQoCQiIiLigQIlEREREQ8UKImIiIh4oEBJRERExAMFSiIiIiIeKFASERER8aDYB0o2m43x48fTqlUrIiMj6d69O3/88Yfr8e3bt9OzZ08iIiJo1aoVM2bMcDvf4XAwceJEYmJiaNSoEb1792b//v2X+2WIiIh4zWQy4utrwmQq9h/TJV6xb4HJkyezYMECXn/9dRYtWkTNmjXp27cvx44dIyEhgUcffZTq1auzYMECnn76aSZMmMCCBQtc50+aNIm5c+fy+uuvM2/ePAwGA3379iUjI6MIX5WIiIg7gwH8zCbMIQbSTAkcsOwhzZSAOdiAn9mEwVDUNbw6+RR1BfKzcuVKOnbsSPPmzQF44YUXmD9/Pps3b2bfvn34+fkxcuRIfHx8qFWrFvv372f69Oncc889ZGRk8OGHHzJkyBBatmwJwPjx44mJiWHFihV06NChKF+aiIgIACaTAd9AWLJrMXFr4tgcv9n1WER4BLFNY+lYpxO2NLDbnUVX0atQse9RCg0N5ccff+TQoUPY7XbmzZuHn58f1113HRs2bCA6Ohofn7PxXpMmTdi7dy+nTp1ix44dpKam0qRJE9fjISEhXH/99axfv74oXo6IiIgbgwF8AyH2+yF0X9jdLUgC2By/me4LuzN0ZSy+gahn6TIr9j1Kw4cPZ9CgQbRu3RqTyYTRaGTChAlUrVqV+Ph46tat61a+QoUKABw5coT4+HgAKlWqlKvM0aNHC7Wevr4mypcPLtRrXg5XYp0lb2rLkkNtWXJ405aZjkzmb5vPlI1Tzltu8obJtKjWgnuvvxcfY7H/+C4xin2P0p49ewgJCeH9999n3rx5dOvWjaFDh7Jjxw4sFgt+fn5u5f39/QGwWq2kp6cD5FnGarVenhcgIiJyHtZMK2NWjfGqbNzqOKyZ+vy6nIp1SHr48GGGDBnCrFmziIqKAqBBgwbs3r2bd999F7PZnCspOzsACgwMxGw2A5CRkeH6/+wyAQEBhVpXm81OYmJaoV7zUsr+lnPixJkirolcLLVlyaG2LDm8bUuTyUiaKYEtx7Z4dd1N8ZtISE/AmgJ2u+Oi63k1Cw0NxNfXlG+5Yt2jtHXrVmw2Gw0aNHA73qhRI/bt20d4eDjHjx93eyz754oVK7qG3PIqEx4efglrLiIikj+j0cDp9IQCnZNgScRoVKLS5VKsA6XsQOeff/5xO75z506qVatGdHQ0GzduxG63ux5bu3YtNWrUoGzZstSvX5+goCDWrVvnejw5OZlt27a5eqhERESKisPhpExAWIHOCTOH4nBo5tvlUqwDpYYNGxIVFcXQoUP57bff2LdvH++88w5r167l8ccf55577iElJYXhw4eze/duFi5cyOzZs+nXrx+QlZvUs2dPxo0bx8qVK9mxYweDBg0iPDyctm3bFvGrExGRq53d7qC0OZSI8AivykeGR1LaHKpht8vI4HQ6i3VYmpSUxDvvvMNPP/1EUlISdevWZfDgwdx8881A1vDcqFGj2LZtG+XLl6d379707NnTdb7dbuftt99m4cKFWCwWoqOjeeWVV7j22msLtZ7KUZKiorYsOdSWJUdB2tLPbOKbfYvpvrB7vmXndJtDu+qdybDY8y0r5+dtjlKxD5SuFAqUpKioLUsOtWXJUZC2NBjAP8hA7PdDzrtEwICoAYxpHYc1xYk+uS9eiUjmFhERKemcTrClQVybsczpNifXMFxEeARzus1hTOs4bGkoSLrMivXyACIiIlcDu92JIwXaVe9Mh7odSbIkkmBJJMwcSmlzKI4MI9YUu4KkIqBASUREpBhwOsnKPbJAoCmMIP8yOOxOLMkOQDlJRUWBkoiISDFjtzuwKzYqFpSjJCIiIuKBAiURERERDxQoiYiIiHigQElERETEAwVKIiIiIh4oUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIiIi4oECJREREREPFCiJiIiIeKBASURERMQDBUoiIiIiHihQEhEREfFAgZKIiIiIBwqURERERDxQoCQiIiLigU9RV0BERORSMpmMGI0GHA4ndrujqKsjVxj1KImISIljMICf2YQ5xECaKYEDlj2kmRIwBxvwM5swGIq6hnKlUI+SiIiUKCaTAd9AWLJrMXFr4tgcv9n1WER4BLFNY+lYpxO2NLDbnUVXUbkiqEdJRERKDIMBfAMh9vshdF/Y3S1IAtgcv5nuC7szdGUsvoGoZ0nypUBJRERKDF9/E0t2fc2UjVPOW27yhsks3bUEX3/TZaqZXKkUKImISIlh9HMwZvUYr8rGrYnD6Kfkbjk/BUoiIlIimExGkiyJbDm2xavym+I3kWRJxGTSR6F4pr8OEREpEYxGA6fTEwp0ToIlEaNRiUrimQIlEREpERwOJ2UCwgp0Tpg5FIdDM9/EMwVKIiJSItjtDkqbQ4kIj/CqfGR4JKXNoVqEUs5LgZKIiJQYjgwjsU1jvSob2zQWR4Y+BuX89BciIiIlhs1qp2OdTvRv3P+85QZEDaBDnY7YrPbLVDO5UilQEhGREsPpBFsaxLUZy5xuc3INw0WERzCn2xzGtI7DlpZVXuR8tIWJiIiUKHa7E0cKtKvemQ51O5JkSSTBkkiYOZTS5lAcGUasKXYFSeIVBUoiIlLiOJ2QYbGDBQJNYQT5l8Fhd2JJdgAabhPvXdDQW0ZGBlOmTGH//v0ADB8+nMjISPr06UNCQsHWsBAREbmU7HYHNptds9vkglxQoDRu3DhmzpxJSkoKq1ev5ssvv6Rfv36kpKQQFxdX2HUUERERKRIXFCgtW7aMt99+mxtuuIGVK1dy8803079/f1566SV++umnQq6iiIiISNG4oEApMTGRWrVqAbB69WqaNWsGQFhYGBaLpfBqJyIiIlKELiiZu2rVqvz555+cPn2a/fv3ExMTA8D333/PtddeW6gVFBERESkqFxQoPfbYYwwePBij0UiTJk2oX78+77//Pu+//z5vvPFGYddRREREpEhcUKB09913U79+fQ4dOkSLFi0AaNCgAR988AFNmzYt1AqKiIiIFJULXkepfv361K9f3/VzdsAkIiIiUlJ4HSg99NBDXl/0o48+uqDKiIiIiBQnXgdKlStXvpT1OK9FixYxbdo0Dh48SNWqVXnqqado164dANu3b2fUqFH89ddfhIaG0qtXL/r06eM61+Fw8N577zF//nySk5Np3LgxI0aMoFq1akX1ckREROQK4XWgNHr06EtZD48WL17MsGHDGDp0KK1atWLJkiUMHjyY8PBwqlevzqOPPkqbNm149dVX2bx5M6+++iqhoaHcc889AEyaNIm5c+cyevRoKlasyNixY+nbty9LlizBz8+vSF6TiIiIXBkMTueFbQt4+vRp9u7di8ORtSS80+kkIyODLVu28OSTTxZK5ZxOJ61bt+bOO+9k6NChruN9+vTh5ptvBuDTTz/lhx9+wMcnK+Z7++23+e6771i2bBkZGRk0adKEIUOG8OCDDwKQnJxMTEwMb7zxBh06dCiUegLYbHYSE9MK7XqXWvnywQCcOHGmiGsiF0ttWXKoLUsOtWXxFxoaiK+vKd9yF5TMvXTpUoYNG4bVasVgMOB0OjEYDEDWEF1hBUr//vsvhw8fplOnTm7HZ8yYAUDfvn2Jjo52BUkATZo0YerUqZw6dYrDhw+TmppKkyZNXI+HhIRw/fXXs379+kINlHx9Ta43xpXkSqyz5E1tWXKoLUsOteWV74JW5p4yZQodO3Zk2bJlBAcH88UXX/D+++9ToUIFnn766UKr3L59+wBIS0ujT58+3Hrrrdx333388MMPAMTHxxMeHu52ToUKFQA4cuQI8fHxAFSqVClXmaNHjxZaPUVERKRkuqAepX379jFhwgSqV6/Oddddx+nTp7n99tvJzMxkypQpdOnSpVAql5KSAsDQoUN56qmneP7551m+fDlPPPEEM2fOxGKx5Moz8vf3B8BqtZKeng6QZ5mkpKRCqWM2Db1JUVFblhxqy5JDbVn8XdKhN39/f3x9fQGoXr06u3btokWLFtx4443s37//Qi6Zp+zn6NOnD127dgXguuuuY9u2bcycOROz2UxGRobbOVarFYDAwEDMZjMAGRkZrv/PLhMQEFBo9RQREZGS6YKG3ho2bMjcuXMBqF27NqtXrwZg9+7druCmMGQPq9WtW9fteO3atTl06BDh4eEcP37c7bHsnytWrOgacsurzLlDdiIiIiLnuqBA6cknn+TTTz9lxowZdOzYkb///psOHTrw7LPP0qZNm0Kr3PXXX0+pUqXYsmWL2/GdO3dStWpVoqOj2bhxI3a73fXY2rVrqVGjBmXLlqV+/foEBQWxbt061+PJycls27aNqKioQquniIiIlEwXNPTWuHFjli9fTkZGBmFhYXz22WfMmTOHSpUq0atXr0KrnNls5rHHHuP999+nYsWKNGzYkKVLl7J69WpmzZpF7dq1+eCDDxg+fDiPPfYYW7duZfbs2bz66qtAVm5Sz549GTduHGXKlKFy5cqMHTuW8PBw2rZtW2j1FBERkZLpgtdRupxmzpzJJ598wrFjx6hVqxZPP/20q+dq69atjBo1im3btlG+fHl69+5Nz549Xefa7XbefvttFi5ciMViITo6mldeeYVrr722UOuoZG4pKmrLkkNtWXKoLYs/b5O5LyhQym/ft6txrzcFSlJU1JYlh9qy5FBbFn+XdNbbufu+2Ww2Dhw4wM6dO3nkkUcu5JIiIiIixc4FBUqe9n2bOHEip06duqgKiYiIiBQXFzTrzZOuXbvy7bffFuYlRURERIpMoQZKu3fv5grIDRcRERHxygUNvb344ou5jp05c4bVq1dz1113XXSlRERERIqDCwqUDh06lOuYn58fffr04dFHH73oSomIiIgUBxcUKH388ceFXQ8RERGRYsfrQOnIkSNeX/Saa665oMqIiIiIFCdeB0q33347BoPBq7Lbt2+/4AqJiIgUJpPJiNFowOFwYrc7iro6coXxOlDKudr2P//8w3vvvccTTzxBZGQkvr6+bN26lffff58nnnjiklRURETEWwYD+PqbMPo5SLIkcDo9gTIBYZQODMVhM2Kz2tEkbfGG14HSzTff7Pr/N998k9dff91tY9nrrruOChUqEBcXxwMPPFC4tRQREfGSyWTANxCW7FpM3Jo4Nsdvdj0WER5BbNNYOtbphC0N7HZFS3J+F7SO0p49e6hdu3au41WrVuXo0aMXXSkREZELYTCAbyDEfj+E7gu7uwVJAJvjN9N9YXeGrozFNzCrvMj5XFCgVK9ePT766CO3xSUzMzOZOnUqDRo0KLTKiYiIFISvv4klu75mysYp5y03ecNklu5agq9//puiytXtgpYHiI2NpU+fPvz6669cf/31OJ1O/vzzT9LT05k9e3Zh11FERMQrRj8HY1aP8aps3Jo4OtTtCJZLXCm5ol1Qj1JUVBRLliyhXbt2ZGRkYLPZ6Nq1K19//TX169cv7DqKiIjky2QykmRJZMuxLV6V3xS/iSRLIiZToe7mJSXMBfUoAVSpUoXnnnuuMOsiIiJywYxGA6fTEwp0ToIlkSD/Mtjtl6hScsXzOlB66KGHeO+99wgJCeGhhx46b9mcSwmIiIhcDg6HkzIBYQU6J8wcikMz3+Q8vA6UKleujNFodP2/iIhIcWK3OyhdKpSI8Ihcs93yEhkeSWlzKJZkLUIpnhmcTi25VRhsNjuJiWlFXQ2vlS8fDMCJE2eKuCZysdSWJYfa8uL5mU18s28x3Rd2z7fsnG5zaFe9MxmWwh93U1sWf6Ghgfj65j/r8YIz2P744w9Onz4NwKJFi+jXrx9Tp05FcZeIiBQVm9VOxzqd6N+4/3nLDYgaQIc6HbFZlZwk53dBgdLcuXPp0aMH//zzDzt37uTFF1/EZrMxc+ZM3n///cKuo4iIiFecTrClQVybsczpNoeI8Ai3xyPCI5jTbQ5jWsdhS0PbmEi+LmjW2+zZs3nppZe49dZbmTBhAnXq1OHDDz/kl19+YeTIkTz11FOFXU8RERGv2O1OHCnQrnpnOtTtSJIlkQRLImHmUEqbQ3FkGLGmaK838c4FBUqHDh3i9ttvB2D16tW0aNECgNq1a3Py5MnCq52IiMgFcDrJyj2yQKApjCD/Mjjszv8StzXcJt67oKG3smXLcvz4cU6ePMlff/1Fs2bNANixYwflypUr1AqKiIhcDLvdgc1mx27X7DYpuAvqUerQoQPPP/88AQEBhIeHc/PNN/PNN9/w2muvce+99xZ2HUVERESKxAUFSs899xzh4eEcPHiQHj16YDKZOHXqFP/73/8YOHBgYddRREREpEhoHaVConWUpKioLUsOtWXJobYs/i75Oko///wzDz30EM2bN+fw4cNMnDiRRYsWXejlRERERIqdCwqUVq9ezVNPPcU111xDcnIyDocDu93OsGHDWLBgQWHXUURERKRIXFCg9O677/Lcc8/x5ptvYjJldVsNGjSI5557jpkzZxZqBUVERESKygUFSv/8849rHaWc7rjjDg4ePHjRlRIREREpDi4oUAoODubYsWO5ju/atYvSpUtfdKVEREREioMLCpQ6derEqFGj+PvvvzEYDKSmpvLzzz/z2muv0b59+8Kuo4iIiEiRuKB1lJ599lni4+O55557AOjatStOp5NWrVoxaNCgQq2giIiISFG5qHWU9u/fz/bt23E4HNStW5drrrmGd955h2HDhhVmHa8IWkdJiorasuRQW5Ycasvir9DXUbJarbz++us0adKEZs2aERcXR5UqVbjrrrto37498fHxdOzYkU8++eSiKi4iIiJSXHg99PbWW28xd+5cOnfujJ+fH3PnziU4OJh+/frx+uuvM2fOHKpWrcrs2bMvZX1FRERELhuvA6Xvv/+e4cOH8+CDDwLQqlUrRo0axdGjR/niiy/o3bs3zzzzDH5+fpessiIiIiKXk9eB0okTJ2jevLnr55iYGA4fPsyKFSuYOXMmt9xyyyWpoIiIiEhR8TpHyWazERgY6PrZZDLh7+/Piy++qCBJRERESqQL3hQ3W6NGjQqjHiIiIiLFToECJYPBkPsCxouOtURERESKpQItOPn666/j7+/v+tlmszF27FhKlSrlVm706NGFUzsRERGRIuR1oBQdHc2JEyfcjkVGRpKQkEBCQkKhV0xERESkqHkdKH388ceXsh752rt3L926dePll1+mW7duAGzfvp1Ro0bx119/ERoaSq9evejTp4/rHIfDwXvvvcf8+fNJTk6mcePGjBgxgmrVqhXVyxAREZEryBWRYGSz2Xj++edJSzu7RUhCQgKPPvoo1atXZ8GCBTz99NNMmDCBBQsWuMpMmjSJuXPn8vrrrzNv3jwMBgN9+/YlIyOjKF6GiIiIXGGuiEDp3XffzZUH9fnnn+Pn58fIkSOpVasW99xzD4888gjTp08HICMjgw8//JCnn36ali1bUr9+fcaPH8+xY8dYsWJFUbwMERERucIU+0Bp/fr1zJs3jzFjxrgd37BhA9HR0fj4nB09bNKkCXv37uXUqVPs2LGD1NRUmjRp4no8JCSE66+/nvXr11+2+ouIiMiVq0Cz3i635ORkYmNjeemll6hUqZLbY/Hx8dStW9ftWIUKFQA4cuQI8fHxALnOq1ChAkePHi30uvr6mly7RV9JrsQ6S97UliWH2rLkUFte+Yp1j9LIkSOJiIigU6dOuR6zWCy59pXLXrrAarWSnp4OkGcZq9V6iWosIiIiJUmx7VFatGgRGzZs4Ouvv87zcbPZnCspOzsACgwMxGw2A1m5Stn/n10mICCg0Otrs9lJTEzLv2Axkf0t58SJM0VcE7lYasuSQ21Zcqgti7/Q0EB8fU35liu2gdKCBQs4deoUrVq1cjs+YsQIZsyYwTXXXMPx48fdHsv+uWLFimRmZrqOVa1a1a1M/fr1L23lRUREpEQotoHSuHHjsFgsbsfuuOMOBg4cSPv27Vm6dClz587FbrdjMmVFhGvXrqVGjRqULVuW4OBggoKCWLdunStQSk5OZtu2bfTs2fOyvx4RERG58hTbQKlixYp5Hi9btiyVK1fmnnvu4YMPPmD48OE89thjbN26ldmzZ/Pqq68CWblJPXv2ZNy4cZQpU4bKlSszduxYwsPDadu27eV8KSIiInKFKraBUn7Kli3LBx98wKhRo+jatSvly5cnNjaWrl27usoMHDiQzMxMXnrpJSwWC9HR0cyYMSNXgreIiIhIXgxOp9NZ1JUoCZTMLUVFbVlyqC1LDrVl8edtMnexXh5AREREpCgpUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIiIi4oECJREREREPFCiJiIiIeKBASURERMQDBUoiIiIiHihQEhEREfFAgZKIiIiIBwqURERERDxQoCQiIiLigQIlEREREQ8UKImIiIh4oEBJRERExAMFSiIiIiIe+BR1BURERK5GJpMRo9GAw+HEbncUdXXEA/UoiYjIFcNkMuLra8JkujI/vgwG8DObMIcYSDMlcMCyhzRTAuZgA35mEwZDUddQzqUeJRERKdYMBvD1N2H0c5BkSeB0egJlAsIoHRiKw2bEZrXjdBZ1LfNnMhnwDYQluxYTtyaOzfGbXY9FhEcQ2zSWjnU6YUsDu/0KeEFXCQVKIiJSbJWU4MJgAN9AiP1+CFM2Tsn1+Ob4zXRf2J0BUQMY0zoORwpXRPB3Nbgy+y5FRKTEyxlcdF/Y3S1IgrPBxdCVsfgGUqyHrXz9TSzZ9XWeQVJOkzdMZumuJfj6my5TzSQ/CpRERKRYKknBhdHPwZjVY7wqG7cmDqOfkruLCwVKIiJSLJWU4MJkMpJkSWTLsS1eld8Uv4kkS+IVm7Be0qgVRESk2ClJwYXRaOB0ekKBzkmwJGI0FuOxxKtI8fuLEhGRq15JCi4cDidlAsIKdE6YORSHQ9ncxYECJRERKXZKUnBhtzsobQ4lIjzCq/KR4ZGUNodqEcpiQoGSiIgUOyUtuHBkGIltGutV2dimsTgy9PFcXKglRESkWCpJwYXNaqdjnU70b9z/vOUGRA2gQ52O2Kz2y1QzyU/x/asSEZGrWkkKLpxOsKVBXJuxzOk2J1dPWUR4BHO6zWFM6zhsaVpssjgxOJ1qjsJgs9lJTEwr6mp4rXz5YABOnDhTxDWRi6W2LDnUlrllr8y9dNcSxqwZk2tl7qFNh2YFScVsZW5Pbem+HUsiCZZEwsyhlDaH4si4crZjKQlCQwPx9c1/7S1tYSIiIsWW3e7EkQLtqnemQ92OeQYX1pS8gwuTyYjRaMDhcBab3CWnEzIsdrBAoCmMIP8yOOxOLMkOoPj2iF3NFCiJiEixVpDg4kraQNdud2BXbFTsKVASEZErxvmCi5Kyga4UL0rmFhGRK15J2kBXihcFSiIicsUrSRvoSvGiQElERK54JWUDXSl+FCiJiMgVrSRtoCvFj/5KRETkilaSNtCV4keBkoiIXNFK0ga6UvwoUBIRkStaSdtAV4oXBUoiInLFu5I20DWZjPj6mpQjdYVQK4mIyBWvuG+gazCAn9mEOcRAmimBA5Y9pJkSMAcb8DObtK5TMaaVuUVE5IrndIItDeLajKVltZb5bqB7Obcx0YrhVzaD01lcdr25stlsdhIT04q6Gl7TLuUlh9qy5FBbXjz3vd5yb6B7ufZ6y27LkyfP4B9kIPb7IeddDHNA1ADGtI7DmuIsNnvRlXShoYH4+ua/8GixHnpLTEzklVdeoUWLFtx00008+OCDbNiwwfX49u3b6dmzJxEREbRq1YoZM2a4ne9wOJg4cSIxMTE0atSI3r17s3///sv9MkRE5DLJ3kDXkuwk0B5GFf+aBNrDsCQ7ybBc/g1xtWL4la9YB0qDBw9my5YtvP3223zxxRfccMMN9OnThz179pCQkMCjjz5K9erVWbBgAU8//TQTJkxgwYIFrvMnTZrE3Llzef3115k3bx4Gg4G+ffuSkZFRhK9KREQuB7vdgc1mL9LZbVox/MpXbHOU9u/fz+rVq5kzZw433XQTAMOHD+eXX35hyZIlmM1m/Pz8GDlyJD4+PtSqVYv9+/czffp07rnnHjIyMvjwww8ZMmQILVu2BGD8+PHExMSwYsUKOnToUJQvT0RErgIXsmJ4oCnssgZ3JpMRo9GAw+HUkgl5KLaBUlhYGNOmTePGG290HTMYDDidTpKSkvjrr7+Ijo7Gx+fsS2jSpAlTp07l1KlTHD58mNTUVJo0aeJ6PCQkhOuvv57169cXeqDk62tyjUlfSa7EOkve1JYlh9qy5Ei0JhaofJI1iWsrXHtpKnOOTEcm1kwriZZEEiwJhJnDCA0Kxd/HHx9jsQ0PLrtiO/QWEhJCy5Yt8fPzcx379ttvOXDgAM2bNyc+Pp7w8HC3cypUqADAkSNHiI+PB6BSpUq5yhw9evQS115ERATCzAVbMTzUHHppKnKO1IxU5m+bT/OZzbl2/LU0mNyAa8dfS/OZzZm/bT6pGamXpR5XgismZNy4cSPDhg2jdevW3H777YwePdotiALw9/cHwGq1kp6eDpBnmaSkpEKvn2a9SVFRW5YcasuSI7stQ/xLExEe4bYkgCeR4ZGE+Je+pO1vMHDeWXib4zfTfUH3q2IWXomY9Zbt+++/p0+fPjRs2JC3334bALPZnCsp22q1AhAYGIjZbAbIs0xAQMBlqLWIiFztituK4ZqFV3DFPlD65JNPePrpp2nRogXTp093BUDh4eEcP37crWz2zxUrVnQNueVV5twhOxERkUuhuK0Yrll4BVesA6XPPvuM1157jR49evDOO++4DaNFR0ezceNG7Pazf1Rr166lRo0alC1blvr16xMUFMS6detcjycnJ7Nt2zaioqIu6+sQEZGrU84Vw+d0m5Nr496I8AjmdJvDmNZxl3zFcJPJeEGz8K72PemKbY7S3r17eeONN2jbti39+vXj1KlTrsfMZjP33HMPH3zwAcOHD+exxx5j69atzJ49m1dffRXIyk3q2bMn48aNo0yZMlSuXJmxY8cSHh5O27Zti+pliYjIVcZud+JIgXbVO9Ohbsc8Vwy3plz6xTCNRgOn0xMKdE6CJZEg/zLYL+/WeMVKsQ2Uli9fjs1mY8WKFaxYscLtsa5du/Lmm2/ywQcfMGrUKLp27Ur58uWJjY2la9eurnIDBw4kMzOTl156CYvFQnR0NDNmzMiV4C0iInIpZa8YjgUCTWEE+ZfBYXdiSXYAlycKcTiclAko2Cy8MHMojqt8/znt9VZINOtNiorasuRQW5YcxbUtzSEGWsyK8XoW3s+P/IIluWSGCSVq1puIiFw9TCYjvr6mKyY35kqqb3GbhXcl0G9ARESKnMEAfmYT5hADaaYEDlj2kGZKwBxswM9swmAo6hq6y6++xVVxm4V3JSi2OUoiInJ1MJkM+AbCkl2LiVsT5zYsFBEeQWzTWDrW6YQtLSsxuqidr76NKzXmtdteo5W5FQG+xW/Nvpyz8FpWa8mYNWNy/b6HNh2aFSRd4ll4VwrlKBUS5ShJUVFblhxXY1vmt1J0tuKyUrSn+oaZwxgQPYB+jfuRZksj2ZrMNcHXUNo/a1abzXrpZ7UVhMGQtfik0c+R5yy84lbfS8HbHCX1KImISJHJWil6sVcrRbeo2oJ21TtnzR4rInnVt3aZ2izrsYxf9v9Cl7ldin2PGBSPWXhXCuUoiYhIkSnuK0Wfm6h9bn3DzGEs77mcuDVx9P6qd67ZZJvjN9N9YXeGrozFtxT4+RW/pG+73YHNZsdu1yrceSlerSUiIleN4rpStMdE7RADNnsGB5IOuMoOiB7Az/t+ZtrGaee9ZvbeaSkkFOskdclNgZKIiBSJC10p2mi8dNGFyWTAP8jAN/sW02JWDFXeqUKjqQ2p8k4VWsyKYdmeZazvu57aZWpjMpjo17gfE9ZN8OracavjsGOn2oRqtJgdwzf7FuMfZMBkUrRUnClQEhGRInHBK0U7Lk2ej8EAvoEQ+/0Qui/snucwWo+FPRi7ZizLeizj1mtvJc2WVqAesTRbGjdWuNF9SC4Q9SwVYwqURESkSNjtDkqbQ3NtFOtJZHgkpc2hHnNpLnbhx6xE7a/zTSyfunEqvx74lV6NenEy7WSBniPJkkRkeCQmQ9Zsq+whOV//4rv20tVOgZKIiBSZi10pujAXqixIYvnEdRPpVLcT5QLLef8EQKXgSoxoOYI9A/cwLGYYYeawIklSF+8pUBIRkSJzMStFnzefqIA5QBeUWG5NItgvuEA9Yk6nk9rv1ubueXdTO6w26/uu50zGmcuSpC4XRq0iIiJFJudK0XO6zckVdESERzCn2xzGtI5zWynam3yiguQAGY0GEiwFSyxPt6Wz9uBaBt480KvyA28ZyNSNU3E4HWyO30zvr3q78p2SrWcuaZK6XDgtOCkiIkXKbnfiSIF21TvToW7HXCtFk2nCbnEC2f8Kb6HK7BWqffwhzFiwxPLygeWpW7Yedmcmvx/+/bx16de4HzFVYxi8fLDb8akbp9Lk2iZ0rtsZh6V4LEYp7tSjJCIiRS57pWhLspNAexhVzTUp7VMWgKTMU7lyjwpjocqcQ3czN39IoG9ggRPL05LsONJ9GNvWc4/YzC4zGdJ0CHd9eleevVYT103Ez8dPCz4WU+pREhGRYsaJ0d/Jkl1fe9wkt0PdDqTaUr26WvZClYGmMFcwknPobvof0/n3mX/5dOunDLx5IL2/6p3vNWObxWJ0mDAYMrHZ7GRmntsjloCvyZcAnwCmbpzK4OWDPQ7tbYrfRGK6e/2k+FCPkoiIFBve5h698P0LLOuxjDCzd8Nl5y5UmXMpgBsq3ECaLY0RP42gRbUWPN748fNeq1/jftx67a0s2/uNK1k8Z49YiKEspfxKcf8X91NzYk1Grxqdb/7TpV5IUy6cAiURESk2vF3LaPKGyfx64Ff6R51/tpzJYKJhxYZULFUBQ46M7pxDdyH+IZxMO0mCJYG7Pr2L2KaxzOwy87zDaG0+bsO98+/NM1k8M9OBv8mfrce24nB610N0KRfSlIujQElERC4LbxaELOhaRv2j+mM05L5emDmMYTHD+PeZf5l/33yOpR7jDKcwBxswB/qQbE1yLQWQbE12rYe0+/RuoqdHEx4UzsqHVrL9ye2senQV25/czqL7F7Hz1E6ip0ez+/RuIO8FI+12B6GFuJCmFC3lKImIyCWTPavM6OcgyZLA6fQEygSEUTowFIfNiM1qd035z1rLKOGCtgTZemyr63jtMrVZ1mMZv+z/hS5zu+SZ49S+Tntql6nN7tO7+fv4365E7s3xm0m2JnN9+etp81Eb7E47If4hJFuT+ev4X3n2EMWtiaND3Y5gOXvM1+TLkKZD6LGwR76v4/mmz2Ny+gA2r163XF4KlERE5JIwmQz4BsKSXYs9JmV3rNMJW1rWEgHebJJrMpi4ocINruDljPUMkeGRrp+PJB9hec/ljFk9hmkbp+U6PzvHaUDUAJb1WEb09GgSLAlM3TjVlcidnbO0KX6TV6/z3GRxk8lAhj2DO2rdweONH8+zHtkGRA2gfe32JGScIiS4dK7gUYqeAiURESl0OZOy88o3yhmwjGkdhyPl/JvkhpnDGBA9gH6N+5FmS+Nk2knKBZbjmuBrePvOt9lzeg/B/sGUCyzHt7u+PW9wAllDZk2ubUL/qP6MXjWayesns77veh5v/DjbTmwr8B5uCZZEgvzL4HCAXykY8t0Qvt/7vWvz3AnrJuQKFAc1GUSzKs2I/iBrKC+v4PFcJpMRo9GAw+HUUN1lYnA6FbcWBpvNTmJiWlFXw2vlywcDcOLEmSKuiVwstWXJUZLa0s9s4pt9i+m+sHu+Zed0m+NaENIcYqDFrBi3oCLnUNrE3yfmCjgG3jyQFtVa0OGzDqx8aCUdPuvg1fBdZHgkix5YRI0JNXA4Ha7n+ev4X1xX/jrqvVfP69d78NmDBNrDMPka3F53mDmM/lH96R/Vn3RbOla7lVK+pfAx+jB141SmbJiSa0ZcdvBoTXHidJ47fJl4dvjSP/fwpXgvNDQQX9/8NyNWMreIiBS6C10Q0mlz3yQ3zBzG8p7LiVsTR++veue5XED2ViDLey4n1ZZaoBwnm93GjRVuBM4mcq8/sp5KQZUKnIxt8jVg8nd/3QmWBEavGk2NCTX47t/vOJR0iG6fdzvvsgE5E8QLcz87uTAKlEREpFBd0AazlkT8/EwYfBy0q9POtUnugOgB/Lzv53yH0qZunMrP+3/Gx1CwjJJUWyqjW492BUUJlgRG/TqK2Vtm8/ytz3t1jdhmsfgYTaw/voYjZ47k+boNGOhSrwvDfhjmWjYge+mC5lWb07BiQ0yGs70b2cFjYe5nJxdGOUoiIlKovEnKPleCJZEyYWVJsp5hzcE1vND8BZpc24Q2NdvQ4bMOXl3jnd/e4e76d2M0GF2BSM7E77+P/43d6b7fm7/Jn92nd7P4gcWk2dI4lXaKsoFlKeVbCl+TL/0a92Pqxqken3NA1AA61u3Ibwd/o265uhxPPZ5nuewE8S3HtnjMtwrwCWDaH9OYvH7yf71dGSzfs/yi97OTi6NASURECtX5krI9BS9h5lAy7FZG/jySaRunEWYOY2SrkTicjgL2TCXxcMOHqRRS6byBSIIlgcjwSMw+ZgYtH8Sg5YO4scKNbksBtKzWkq8e/IpmVZrx9m9v58qNeuaWZ4ipGkPc6jieb/o8cavjGNJ0CGHmsFxDatmLWua3dMHAmweyvu963vv9Pcw+Zr7c8SUmgylXgHeu7CUKTDYlexc2JXMXEiVzS1FRW5YcJaktz03KPl8vytc7v+aRiEdYuG0hDy9+2HWN5lWbM7r1aGJmxnj9vFv7b6VySGUW71icK/G7caXGvNTiJaKviab/0v7cf/39bDu5jdGrRnu81vaT29kcv5n+Uf3depyy93DLTsb+pNsnlDWXJdGayPHU4zyz7Bm3azWs2JD5983Hx+jjcemC7N/RM7c8Q4Y9g9Ppp/Ez+eUK8Dw5OOggvgZfjqUeV7K3F7xN5laPkoiIFDpHRlZSdveF3fPtRXnu1ucA+GL7F27XyLlitjfCzGFUK12Nod8PdRuuOjdIO3zmMJPaTyIsIIyGCQ1ZfXA1v+z/xa1OLzZ9kRphNej1ZS+2HNvCmNVjcvU45Vx88q01b/HzIz/T+qPWLOu5jEHLB7k9/vfxvykbUJYlO5fkGSTl/B3d+cmdHnua7vr0Lteq4Oc6lHyIISuGsOrAKtd5+S03IPlToCQiIoXOZrXTsU4nBjcZzJM3P3neBSB7fdmLAVEDGH/neFYdWOXqNTl3xez8/N9t/8fSXUvdgiRvVule2n0pKdYUTqSdJMwcSmlzKEaHD6fTT7iG/RxOh9vq3+faFL+JREsi1UKrkZqRysiWI3nlp1fcyjicDsb/Nj7Xudkz+873O+r9VW/6Ne7ntkjmuULNoSRbk93OO3etKvUsFZxmvYmISKFzOsGWBq/d9hprDq7xagHIXw/8yoCoAa6ZYDdUuIFpG6cx8OaB+T6fyWCie4PublPzvVlaoPvC7sSuiCXIL4iq5poE2sOwJDtxZMLp9NMFes2n009ToVQFDiUfondkb7eNdW+ocAOn00/nmW81IHoAv+z/hWkbp513JtzUjVM9bgQcGR5JgE8Afx3/K9djee1HJ95Tj5KIiBQ61wrSTgfj1ozz6pyJ6yby8yM/06tRL1cOU/bss/y2AhnRcgSpGe5rKHm7tMDkDZNpUa0Fnet2BqsBpxNMPlA2sKx3L/Y/ZQLKcDz1OGEBYTzwxQPEVItxzaaz2W1YMi25zjEZTPRv3J8eC3swLGZYvgnoE9dNZNEDixizeozb0N7AWwYydePUPPeig7z3oxPvqEdJREQKhcGQtSK3OcRAmimB45mHSSjgJrdHU45y/xf3EzMzhuvev47Oczuz+sBqxrYdy6wus3ItAhkRHsHMLjPpHdmbg8kHXcdNBhP9GvdjwroJXj133Oo4EiwJfLvvK/yDDfxw4HsMGAq46GRp9ifu55rga2hZvSVTNkyhxoQa3P/F/by19i1CzaGuumX3GnW9ritOnMzsMpPaYbXpMrcL171/nev13z3vbmqH1WZ93/XULlPbbSPgbP0a9yOmagxTNnheRiB7rSqTSR/7BaUeJRERuWh5bYCbPWutIE6kniDEP8T18+b4zdw7/14GNRnEC81f4PYat2N32kmxphBiDiHUHEpqRion005SJaSK67yc6xZ5Y1P8Js5knGH06tH8vP9nnm/6PDM3z3RtlJuf55o+x5oDa3jq5qeYtXkWtcJquZKvtx7bSrMqzQgLCGNiu4l0qdfF1WtUrXQ1QvxDiP0+1uv8pFNppwjxD3FbouCuT+/KM28p53IM6ZkWgkwG7FfQUkvFYW87BUoiInJRPG2AW9BZa5A13JWakUrDig3d1loa/9t4GlZsSKh/KG1r3UGFwAr0XtybY2nHeCLqCVpUa0GAb4Ar8Tt73aKCyA5Apmycwi3X3oLT6aRFtRb5DvsNiBpAp7qdGLt6LI/d9Bg9F/YEQ9YstO97fc+GIxuIviYaS6aFMHOYW1L5hLsmUC6wnFcrj2dv4lu1dFXm3zufQL9A3lz1JoOXD84VJOW1HIPZxx+T2YHBZCrWSwa4722XcHZvu8CiWe5AgZKIiFwUX38Tqw7+zF8n/qJhxYb8ffxvQvxD6FyvM5WDK3s9a61ltZaUCyzHogcW5ZmnM2vzLJZ0X0JmOth9Lbx919uU8i3F0l1LuevTu2hfp72rB+hCg7TsWWPZuUBtP27LN92/4dZrb2XCugm5Zs09e8uzdL2uKztO7OCF5i+QaElkeufprroH+wfTrnY7TqSd4OUfX3YLiEwGE3fXv5vOczp7Vb+J6yby9YNfE+QXhM1uI9GSyLe7v80VJHkz06+4LhmQV89ktqKquxacLCRX0oKTJpORMmVKAXD6dCpGowGDwYDT6dRqrlegkrRI4dXuSmvL7G/+Jn8HR84c4Xjq8azgwC8Yk9HEt7u+JcWWQpBvUL7DV7XL1GZtn7V8u+vbPFfBHnjzQFpUa0GIXwgBvgEs27OMmyvfzKhfR7mCjzBzGOv7riduTRwz/pjBv8/8mytQ8CQyPJIv7/+SmhNruhKitz+5nfu/uJ+DSQfpH9U/16KT1wRfg81u40TqCSqHVGbRjkV51v3Dzh8SnxLPG6vecFuNvGHFhsy7dx7XvX+d17/zvc/spbR/aT7e+jHHUo9xY/kb6b6wu+vxMHMYGx7f4HGpgWzZSwZYU5zFpmfJYAD/IEOunslzFVbdvV1wUoFSISkugVL2eG524JP9X6fTiY+fEZOvkxRrCplkEuwXTJI1iYT0BMLMYYSYQzBhwsfkg8MGNpsDw3+7LNrtjlxBlMlkxMcn6/nAgMPhIDPToUDrMrvSPlzFsyupLc9+8//a9c3fZDBx67W38sX/vmD6xumM/HkkIf4hruDF0wd3mDmMv5/4m//7+f/O+wHZr3E/xt0xjhE/jshaXDK0Gg8tesitTHZvyq8HfuVMxhmvgjSAmV1msvPUTrdVulc9uooXVr7gWsDRaDC6LTr5cdePOZZyjMaVGjNp/SRe/unlXK/riegnGNpsKEdTjroCyewesk1HNzEsZliBVx6vGVaTau9U48YKN7Kk+xKGfHc2sBgWM4zaYbW9es1zus0pVvvD+ZlNfLNvsVvg50lh1F0rc1/hPCWw5XX83PFcq91KucBy+Jl8SUg/jclkIsgviJNpx0hITqBmWE1W7lrpsVuzfZ32pNvTCTIHYTQYSUhP4LQ1a32QEP8QTAYf7M5MMh2ZGA1Zu22n29IpbS5NcEAwRoMPjkwD9kyHK9p3Op3/BVVGV92zg6/s15T9WrJfn9OZdTM2mbLOyciwk5GRednaQETydm5OUpg5zDW13e6wczj5MPfecC+9GvVi2h/T6L6gO5/f9znNqjRj/G/jc913pneazk/7fsp389epG6cSUy2GyiGVeeymx2gxs4Xb4yaDiUDfQJ7+9mm61OvCgzc8iJ+PH78f/j3fACymagyDlw92O55zKA5yLzoZ4h/C38f/xs/Hjy//cd+TLTtgW31wNS1mtcizh+zxmx4n01Gwe1qIfwgLti3gVPop+jfujzPDRFybsbSs1pJxa8fRr3E/us7tmivHK6+94orbkgFGP4fbOljncznrrh6lQlIYPUruAU/i2QQ2cyhGhwmH0e5+3D8U7EYwOfju3+UcSTlC1/pdcTqdHE89TnhQOCH+ISzZuYS4NXHsT9zvVZdsv8b9GNJ0CL2/6s2AxgPoVK8TBkPWeigp1hRSbamU9i+N2ceMzWHD1+RL9p+RLdOGv68/BoOBjMwMDAYDydZkEi2JhAWEEeIfwhnLGTBAaXNpfAw+JKQnYDQaCfQNPNvD9V/ZdFs6viZfAnwC8MEnq4cMcNizgi2jMWvNk8xMByaTAaPRiNPpdAVh2T1ikNUrVhKHGK+kXgg5vyulLXN+869dpjbLey5n9YHVeQ47DW4ymC71u3Aq7RROp5NypcqRZEniWOoxSvuXpmJQRQwYiJkZk2uGWl4b6Das2JCfH/mZ46nHqf1ubQDKBZTj5ZYvc+/195KSkUKyNZk6Zerw7a5vWXtoLc81fY6f9v2UZ5CWc9ZYzq1B8hqKyykyPJKvHvwKh9OBAcN/XxyNfLDpA3ae3MnkDpOZtH4SI38e6XFD236N+xHXNo5Oczq5baHiSWR4JD898hM1J9Tkfzf8zzX8BFmfHT7+YMlMJ9WWet61mHI6+OxBAu1hRX4/NJmMpJkSqPJOlfwL/+di666ht8vsYgOlvLqxs2XvhRRTNYY2H7dxvZmze4Ba12hNWmYaP+79kVlbZtG8anMGRA0gxC/Ebc+jgnTJftrtU8oElOH68tdjd9hJyUg5G+hYs27iQX5B2J12fIw+pNvSCfQNxGQ0cSL1BEnWJEL8Q/Ax+vDlji/5dte3HD5zGF+jL0/f8jS3V7+dOX/OoXGlxtxa9Va2Hd/Gin9XsDl+MydTT/JwxMPcfd3dGDCQaEkk0ZJIqDmUEP8Q7E47fiY/MuwZxJ+JdyVMOpwOEiwJrkAr1D+UTEcmiZZEgv2DCfQNxGjICqSsmVZ8TL744ovD4cRmc2Cz2V29WUCRT0n11pXy4Sr5u1La0hxiIGZmcw4kHWBTv028uerNfHtshjQdQvT0aJKsSdx73b083OhhGlRsgNnHzJmMM9SaWMtV/nwb6E77Yxr9G/fnZNpJWn/UmpdbvEzfxn1Ze3AtH/zxAT/s/YF1fde5fSEMM4cxIGoAg28dTEpGCkdTjhJqDs21sW1OeQ3F5fX4Q40ewpJpYdjKYbSr046eDXqSakvleOpxAn0D893Q9uOuH1OvbD1u/uDmfH/vc+6ZQ50yddifuJ+2Ne/AlgZgcPXImwIcfL3zK8auGesxx+vcgHBr/z+p4l8Tm61oh998fU0csOyh0dSGXp9zsXVXoHSZXUyg5G0CW86bTfYbLufY/vd7v3fNdDg3gdJkMPHvM//SeU7nfNcVqV2mNj889AOrDqzKM2jLfsMNXj6YB258gE51O5FiSyEhPYEgvyBMBhOztszC6XTS56Y+WDOtJFmTMPuY8Tf5M+2PaVgzrbzS8hWOpx4nyDeIEP8QUmwpGDAQ6BuI3Wnnm53fsDl+M6HmUIL9g9kcv5mUjBTurHUnt9W4jS3xW7i95u04nA6W7Fziujlk32SfjH4Sh9OBESPB/ln5WKfTT1MmoAyh5lCMGEm0JOLESYh/iFtvVmn/0iRbzoDBSYhfabCbwJG1LYPTmdVmgFtO1rn5YTmDrOx8rnPPKQxXyoer5O9KaMuc3/xfa/UaNcvUpMfCHvmel1fg0a9xP8a2HcvfJ/7m1hm3Au4ztib+PjHP+0/X67pizbRiNBgx+5g5nnqcoylHKRdYjnKB5fjt4G+MWTOGZGsyR5KP8HjU4/Rr3I90WzqQ9SXPgIFJGyYxaf2kXAFM/8b9eb7p8x73VMu+Fz+y6BEmdZhEhVIVsGRa+GHvDx7rnFeQAmd7iWJXxDJ141SPv79+jfvxYvMX2XBkA3fWuhMyTTh9skYZkizJVA2tQuyK2AJ/hqhHSYHSZXMxgVJBEtjOvdlk9xI9991zrmG17NkeOYMib2dXhJnDCvwN8YEbHyCubRwnUk9w+MxhqoRUoWxgWb7+J+/esWdueYa7at+F2cfM6fTTHDlzhAqlKlAusBxmHzMAR88cpVJQJexOOwmWBCyZFkLNoQT5BmF32DmYfJAyAWXw9/Fn5qaZzN46m7+P/02NsBquRM4vt3/J23e+7fGGG9s0lg51O/D9nu+Z9/c8jqQccXXvN6jYgJdiXuKOWndkLW6XkcLp9NOEmkMJM4dhMpiw2W0YjAZ8jb7YHXYSrYkkpCe49XyZTVmvJ/t1JFoSKRNQhtL+oRgyTdgzndjtnnut/Px8XD1bnvKzroQPV/HOldCW2d/8b5oWydHnjtL247ZeLeoYGR7JogcWUWNCDbehrG+6f8N15a+jxoQaBZqxFdc2jm93fcsbq95wfUE6N3G6UlAlypcqz+Idi/McFnz2lmdpXrW5K4CJCI9gaLOhtK/TnhV7VvD6r6/nOmdI0yG0qdmGDHsGKRkpZNozqRZa7YKClGz/PPUPAT4BrNy7Ms8lCM4dHhwQNYDXbnuNbp9345f9vxRotCDnZ0hkeCQ/P/ILluTiEQaYQwy0mBXj9SzFi627AqXL7GICpexu7ILcbLrM7UJp/9LMuWcOHT/ryF117nK9URpWbMjn935O/ffru87LXiE3v9kVr7V6jVplankM2nLmDDx/6/OsP7KeUb+O4qO7P+LO2nfywBcPsOiBRQxdMTTf6Z1Dmw0lcmpk1s/RA3gi6glsDhtnrGcICwgj0DeQ7Se2Uy20GikZKZxMO+kKqPxN/hw5c4QA3wCC/ILIdGTidDoxGoy89ONLzPtr3nlvuCaDiVur3Mqo20YRWSmSI2eOcDr9dFb3vm8A8/+ez93172bVgVVMWDcBh9NBw4pZXcLWTCtd6nWhU71OGA1Gftr3Ey//+DIbj250XT8iPIL/a/V/tKreioU7FvLOb++4vfkbV2rMa7e9RkzVGCyZFvxNZtdCagYDmAN9cJrsWYn0//WChZnDMDhMWNIyceSIq66ED1fxzpXQltnf/DvP7cziBxZT9Z2qXp+bPd0+Z0J0VKUofnzkR2JmxtC+TnuvP/A/uvsjtp/czuhVo129UKsOrOKdde+4Aidvg643Wr/BqbRTrvSBz7Z+xsn0kzze+HG35QByLnuQ/eXrQoOUnFY9uopRv45iTJsxVA6pzPHU467n9DQ8mH2tuNVxXo8WgHvA+mnXT887c+xyr4qtWW9FyOFw8N577zF//nySk5Np3LgxI0aMoFq1akVdNUwmI0le7oUUZg6jXZ12hJpDmXfvPE6lnSLTkcnC+xcS5BdE13ldgaxZEckZyW7nerP4mslgon9Uf9p83CbP5z43Z6BCqQrcVuM2AGZsmkGHuh348v4vWbxjcb6zVyZvmMzNlW9meMxw7q5/N7/s/4WOczrmCiZGthpJbWNt3l//Pn8e+5MESwK+Rl+eueUZutTvwidbP6Hbdd2wZlpJtiYT4h/CC81e4O56d7P24NpcN8icw3LZC9U9s/yZPHubSptLY8208m2Pb7FkWjidfpqygWUJM4fhcDo4mXaSJGsSNcNqsuB/C5j2xzSmbZjGNSHXUCmoEpGVInlx5YusObiGRhUb0ahiI/Yl7COmegx9b+pLmi2NbSe3USmoEhmODI6nHKde2frYyWTRPwsZt2Zcnrlqd9e/m8w0AzZb8c6dkpLJ4XAQGhRKp7qdOJ1+ukDnZq98ndOGoxvIsGcwtOlQmlZt6vXii+N/G8+iBxYxbeM0lvdcnisgGhA9gF/2/cJvh36jedXmHmd/ZW+Im/16Tqad5J7r78FoMDJlwxR+2v+Tq+zsu2cz5pezz5O9n1xBFozMa0Pb8qXK079xf6qFVqPVrFbYnXZXAvtfx//KM5k8+1rf7v62wFu1pNnS+L9W/0eHOh2xprj/PopyVWyb1U7HOp3o37h/vl+086r7pXJVBEqTJk1i7ty5jB49mooVKzJ27Fj69u3LkiVL8PPzK9K6GY0GTqfnHgM/V85x+5azWubZHTyzy0zu+vQuUjNSCS8V7nb+38f/JtA38Lwr5Das2BBLpiXXGy6/VV6fveVZHo14lCRLEuVLleettW959donrpuY59j8uUHZ/qT9DG02lCC/ICyZFmx2G/O3zSfNlpYVPH7aLs9Ap3nV5tQuU9uVE5D9On479BuZjszz7q3UfWH3XN3b2ed/se0L17fWnM85uMlg9jyzhxOpJ1zf/Ma0GeOWZF42sCxBfkF8/vfnvPD9C65yEeERvBzzMrXL1nZbE+XcevX6shf9G/dn7B1jsZ/BrWfp3G9/xWGPJCl5fP1N7Dr1D21qtCEsIKxA55473T7boaRDdK7fmaNnjhboAz/DnsHUjlPZemwrM/6Y4XqsXEA5BjcZjCXTQpMqTfKc/ZVsTXb1jn+5/UtiqsZQfUJ1V1CSfR958uYn6fhZR+6sfSc/7/uZGX/McE29Dw8Kv6Ag5cYKN7p61SLDIwkPCmf57uUcPXOUTfGbCnStRhUbFXirlnRbOoNuGYQl1eEW9BT1qthOJ9jScC13MGbNmFx1GNp0KB3qdMSWxmVbKLPEB0oZGRl8+OGHDBkyhJYtWwIwfvx4YmJiWLFiBR06dCjS+jkcTsrkc7MJM4fl+Y0p2+b4zfRY2MO1aWKfr/pQ2lzaLSiyO+1M3Tj1vBs8RoZHctri/g3Rm+d+ZPEj9GvcjzfbvEmSJalAN43T6af5/fDvrhtPqDmUd+58x2NQNvDmgbSp2YZHGj3C8B+G5xlQ/HnsT95c/Sa9Gvbih4d+4KapN2F32l2vo1xgOTLsGfnurZTd69WsSjP+PPZnvr+HhxY95MpD6L+kP/PunecxqBrSdAh/9PuDth+35VTaKdrXaU/L6i356p+v8u2Nm7JxCi2qtaBz7btJT81az8qaaSXNdPbbX2ipUDLsVg6fiae0OaTI9kiSksfo5+CZZc/wabdPc91nzicyPJIgvyD+Ov5Xrsf8fPxYtmsZdcrWKVBd0m3p3FnrTlJtqex9Zi9TNk5h+e7lLPjfApbvWe5x9te2J7fhcDpItiaTbkvPyhs0l+adO99hxE8jSLAkuH1h+uHhH0jLSOPHfT9yePBh0mxpxKfEc23ItXkme59Pdq9adhrDG63fYPbm2SzasYioa6KAvJdFyGuJgZSMFEr5lirwVi3lA8tjTXO6BTue9uvLlvP3MaZ1HI6USxOo2O1OHCnQrnpnOtTtSJIlkQRLImHmUEqbQ3FkGLGmXN77WInPUdq6dSv33Xcfy5Yto0aNGq7jDz74IPXq1WPkyJGF8jwWi51Dh7zLUQoJOTtrCrJylJpNbsef8X/mWX5w0+eoGVqTp755MuuAKQP80t0LZfqCLZD3O0zCmmnltuq3sfrg6rPnAKXNofz48I9MWDeB2dvfBaN7L8OQxv9HrwaP0HByA8/Pnc2YCf6pboc+6TyXiDLNuXHyDV79HkJD/Ng04Hf8fPxItiZzKu0U5XyqYXD4MfWPqcz4YwZJlsRc533c9RPSM9N5fEkfMJ/N5Qgzh9G/8RM8WKcf6ZnpnEo7ReWQygT4BLD12FZOpZ1iwNIBbO6/me4LuvPX8T/BNx18MtyfwGaGTH8AGlRsyKfdPmX2ltnUCK2R+/cAEJDk9uPMLjNpXr4jr//6OrM3z/L4+ntH9uHZJs/i42tn3bGfaFurLR0+65AVaGb6gS3A47kNKjbkmx7f4HTC8j3LeG/LG2w5fvabaER4BH2vf45br23GPZ/fQym/Ujxz8zPcWeuurH2yzvk26OMDpUqd82uwQVoB0u5Klcq6Tk4pKXi9U7nJBEFB7sfs9qxreCsgAM7tJE5Ly3ot3jAYst6fOTmdkJy7E8Qjf38wm92PWSxgteZ/brlyWTlKGRln3O4RkFUHb+/Wvr4QGOh+LCMD0tPzLp+XoKCsNskpLc1IConcOPkGapepw0+P/MTifxbz5NIncl/gnHvER3d/ROvqd3L9+Jvd3tcNKjbkk26f0GNhDz7t9imNJueYHu6XCqZzJjFkBILdF4D1j29g1C+jaFOzDTHVYvjr2F80r9acET+OYNbmmWBwuN0jAHAYeLLRMAY1GUS6LZ20zKwcpPKlylPGXAa708598+9jy38BVmiwmZW9v6Fu2bqcTDvJ0TNHCaAMRkcA3+7+ljY12hA1rfH5f5k57hE7ntzBd/9+R6uKd5PpyCTNlkaQXxBlAsqQ6cjkgz8+4IEbH3Ddw8oGlqV0oJmPtk91X2Ig048tj+3CiBGDwUCPhT3IdGQS7B/MGesZtp/YjiNncOWfDEanWxL0mRy90r7+JlYc+Ia+Xz+W92sw2cAv64aQnR+UesZ+ye8ROXvGwVGo9whvc5RwlnDLly931q1b15menu52fODAgc7HH3+80J5nw4bsieP5/0tIcD/XZrc5A4OtXp9Po5lORuL+r8vD3p+P0/nq3MXOiCkRbteoWOuo99eo9mOuOtQe9FiB6jD5oyPOT7d+6l6Pel96f43Se88+98Tazt2ndjvjlnxeoDpw5zO5f5e3jC/YNc45P3JKZIHO79Brl7Ph5IbOf07+c/Y6dz5TsDo8Uy336yi91+vzu3TJ/Tf95ZcFeH6czh9/zH2Nli29P79Ro9znb9pUsDrMnJn7Gg8/7P35pUvnPj8hoWB1GDEi9zVGjCjYNc69RzidWXXz9vyHH859/syZBavDpk25r9GoUQGukeMe0e/rfs7dp3Y7h89YUbC/6/u75P67vsB7RPa/aiNaFqgOL79xyjl782z3+9QF3iMip0Q6kyxJztmbZhfo/Aq3f+r88I8PnbtP7XbWnljbyUicle8bW7A6/HePmPPnHKfNbnM6nU5ntWoFOL/el26vI8WaUmLuEfkp8UNv6f99hTo3F8nf35+kpKS8TrnsfIw++Bqdl/U5UzJSWPnQSlIzUl0LMnZcUJZjF3HN3ad3Faj8lzu+5Lt/8+ihKSC34cHvlwP3XfQ1L4a3OQbZKpaqACFVSM8swNf9QmZ3ZJLpyPpbFCks505tZ28rIPdkEU+qhVbnUI5tQS5WmDmMj+76iJaven/OhsMbeGNxb7c6+PuY8aKDMJfYZrGs/HclDy9+GHgo3/LZOtTtyO6EMfx5/E+W9VhG9PRoWlRrwZwCPLfRYKJf1AA61e100e/zTfGbSLQkAqXyK1oilPi7ovm/PvCMjAzX/wNYrVYCAjwPbVxKJ0+eyTUU4HQG5V34EmlVvRVvrXmLX/b/wshWI0m2JON0XnNZ6/Ddnu/A+02zPRoQPYCf9/38X+5QtYu/4GWW6bTTo0EPSvuXLrI6/HLgV+b/HU+nep1IsaZwPPUEJ1MrAd7nPiQmpnHihPsHms0WgLe3mcxMOydOuPfjnz5tpCA34zNn0jlxwn2oxmIxA75ene90Ojlxwr0fP+v7VLDXdUhNtXLiRMY5x/wAf6+v4fkeYciz/LksFhsnTrhvgnXmjA/g/T3vTGo6Z6zu2yY5nJUAL4YqgOjKNzP//kVM3TiVwcsHZw0Zna7s9fMDPN/0eTp1HnTela0LYkD0ANYdWgd4v6RBo/BGTH5qD4v/Wcyy3cswGU2UO/oAs1YX8LmjBtCmZhvqvlu3YCcCX+/8mriw2rSolrVn3IzOM/jjy5oFusbGfhupW9MPa6qDtKSs4UiHoxRgLHB9AE6nJ5CaWo6C/F0Xt3uE1lH6T3aO0ooVK6ha9eyb48EHH6R+/fqMGDGiUJ7nYnKU4GwOhI+fCZOfI2vbDmuiK+nvxkk58n7Ok6ME8NcTfzPq11EMaTokKx9p8ywMBiPXl7/eNX69LXkdTkOma2ZXnXfrkJhs58OOs7m1yq1ZU9dPbKNeuXpET4vK/SLyyFHCbmJLn38xYmTc2nF55ucYDaazOUKJv503/wDO5gg1mtIIp9PBDRVu5MMuH3LL9JvB4MAUkOa+hojDANZzkkxyaFCxIUu7L6XjnI5sjd/iVY7SnHvmkGpLzXrOvATk7plc1/Mfei/uzd95JK6e66c+31G/UjUyHZlnl0jwIkfpk26fEDElAqfT4co/cGMJBmfWTXBd39/pvbg3BoOB/o3707RKU+75/B72JvybVfa//AO3RfFSz3BjWJPz5jblpBylLFd6jpLJZMAnAH6JX8K4de6zjhqENuWJm56mba22tP24LRWDKvJKi1do/2k7APf7jC2B7cnr3ae2202s6rmF6qHVOZF6AgxQLrAcE9dN5MEbH2TjkY1M+H0Cfx37L1/TL5WIyje67QbQvvp93BzezPX3W9ocSp+b+jAweiDJGckcSj5E2cCymH3MfLR1Fh9ue4cES4Jrd4JOn3Zh64G9+f4eXPeeGXVpVLk+Q5oOoX2d9vgYfBj149v8r14v1hxcw5SNU0jNSGXB/xaw9tBaJm+YzF/H/sRoMHFd+etoVLERD0a1o0X1FvRc2JOvd36d9QTpnr8YnXvfw8cKvhbX6txhAWFM+202pYzleH/9+677ujXTyuf3fe6ql+v3CDSqVpOhzYe4zVpLTTVgMsNvh9dSI6xG3vd61x/G2RwlyFoV22wrS2Ki95v6Frd7hNZR+k/9+vUJCgpi3bp1rkApOTmZbdu20bNnz0J7HpMJSl9Eh8DZm7Mdk8mIb6Av5Q3lmbJhCs80eYaIGjXOP7PExwY+SUSGR1KlYjDz9kxl/ekVLOuxjFb1I/5b7XU1/JfXGFEpq0u8RbUW+Pv4c/8N9zNl4xQ+/3caN9WsSetPW/FE9BN0btQi/+f+T2TlhlQPD2XcmnG8fMfTOZ737Ln3Xv8/gkvb+evMqry/mPq5B5t/Jv+Kb6lUGlSrwtZjW9mR8hvh5cyuOt1QoaH79FyjM8/AJef1bH4nePWuQa51p3LxtWT9Awbf1pvPdk2iX1Q/738P4ZFULOvP9pS1EJD/tPwq5cqRZkvjg00fnJ2V6JORO4DLYVjbJ/gx/kuc5vN8w86RxGrzPUFYmIFVB1bx1I+r6Ne4Hyv6zs+1SvDUjVNpcm0T+kf1Z/Sq0fyVvIq+369iQOKA/zbg9P7DGnLf1ArqYt9XkDtgKCiD4eLrYDbnDp7yEhqa9d8TJ3I/dm4AV1B+frmDyHNlbafkeebTn4lrGPDDGgYkD2DVE0u4/4v7qX1NOdd7zgn8febsfSYXk52K5fzpOL+5a7HGG8vfSP/mDzBm9ZtnZ5Tm+I6wOX5z1gbdUQP4uNvHTPhtAq0/H0iCJYHalc8uXXL3V63ynOG2vtl67vr0LoL9grFkWth6YrNXHWuue881dfnz2J+MWT2Go2eO0juyN/2aPMToVaPdZsC2/rwx/aP6822fOfgYfQjwCXDbEBygQcUGrDqwKus9l899yj8onYdu6cym+E3/zXo7+/7sVLcTTzZ9jNPpp7ntukhOpp2kfGB5KgVXYuW/K3l57dP/PcfZa245vslt1pozFcqGZ7X19D+m8+8z/xboHhfsH8wZwykqVgnFaTOSYbmwmWjF4R6Rnwvrc7uC+Pn50bNnT8aNG8fKlSvZsWMHgwYNIjw8nLZt2xZ19fJktzsI9CvF59s+p1F4IyyZFgbePNCrc2ObxZKQnoDD6WD36d1ET49m56mdLH5gMduf3M6qR1ex/cntLLp/ETtP7SRqWhR7E/byaqtX+bTbpwT7BWP2MZNgSchaKXb1GJ695VmvnnvgLQOZvnE6Dzd6mPVH1pNsTXZ73n+e+ocPOn/g2lTXWzkXqcte5iC7TiH+IQVeQyQ+JZ7oa6J5vPHj5y3Xr3E/YqrGMHnDZNfSCt6IbRbL4n8W57lI3LkiwyMp5VuKw2cOM3l91sJ33tSrXe12LN+93Kv6QO71a6ZunMqvB36lf1T/XGUnrptI/6j+GA1nbw+TN0xm6a4l+Pp7N+wiVyZffxNLdn3t1YKxP+/7mfF3jKdcYDkiwiO8un5keCS+Rl+2HtvqWqzxZNrJHEPn53/OhdsXkp6ZToIlwZWbGLcmjt5f9c71AZ8dYL2//n1W917Ntz2+de315q0kSxJPRD/Bv8/8y7x753HLtbeQZE0i1BxKucByhJnPLu2SYElg/rb5WDOtLN+9nFazW1H1nao0nNKQKuOr0HJWS2qH1WZ93/XULlM73+e2ZFoYfOtg5t07jz0D9zAsZhhh5jC+3P4l/j7+LNqxiI5zOnLd+9cRMzOG+u/Xp+WsliRaEs/7HNnvZXOgj6utcy4f442BtwxkzOoxVBlfhRazYli6dzH+wQZMJu+Ghq80JT5QAhg4cCD33nsvL730Eg8++CAmk4kZM2YU+WKT5+PIMHJN0DU0qNAAo8HI7TVup1/jfuc9p3/j/nSs29EtUS/BksDoVaOpMaEG939xPy+sfIH7v7ifmhNrMnrVaBIsWXuT+Rh96FCnA2/d+RbhQeGuG9+k9ZNoXrV5vh/eA6IG0K1+N1pUa4HdaadNzTY8eOODpNvSSbelU7tMbULNodgcNkr5FSwB8NwP+cnrJ9Ohbgf6Ne7n1Yrj5yoTUIb+S/sT2zSWmV1m5rrJR4RH8Fm3zxjRcgR3fXoXCZaEAgUxHep0oGKpil7VJbZpLN/v/Z5ygeVIsCRw16d3nbdeM7vMZEjTIaTZ0gj28y5vJjI8kgCfgFzr1+QVEIH7ong5xa2Jw+inhStLMqOfgzGrx3hVduyasVQLrYbZx8yQpkO8OmfgLQOZunEqDqeDGyrcQJotjS71uzBh3QSvzs/5N+uem5i32mVq82T0k3y35zsGfjsQfx/v82lql6lN/XL1MZvMdJnbxRWQ1JhQg9tm35Yr6MkZuD2y+BGPgdvYNWNZ1mOZW5CVFz+TH48seoT7v7ifV39+lehrotn4+EYmtpvIkO+G0H1h9wt+jrg1cThNdsauGes6VpB7XEzVGKZsmOJ6zu4LuzP0+1h8A3OnlZQEV0WgZDKZGDJkCGvXrmXTpk1MmzaNa6+9tqirdV42q507at7J7C2zgaw/xthmsXzW7bM8P0Bn3z2bIc2GYM20ulbgzsnhdLD12FZWHVjF1mNbXb0dkeGRlPIrRYIlgajpUVR7pxqzNs/i2SbPAnj14f1x14957tbn6PFlDwZ/N5j75t9HxXEVCX8rnP998T8GLhvI6fTT/PDvDzSd0bTA30DP/ZB/4MYHSMtII7ZZLIObDCbYL7hA1ysfWJ4jZ4649bb989Q//PH4Hxx49gBfP/g1NrsNf5M/r932GhHhEW6/h9l3z/YYxIxsNZIf9/5I1DVRXgWXHet1ZPCywa4286YXsM9XfTAZTbSr3c6r15zzwyknTwER5L3VxKb4TSRZEjGZrorbxlUnazulxAItGJtgSeBY6jE61u2Y7xe5cz9gQ/xDstYxuoCVrRtVbES/xv3OG2DlDFx6fdmLBdsX5Hlv9HTuyodW8sL3L3gd9HgTuMH5e3OzRYZHUiagDIseWMS8e+fRO7I39cvVp7S5NKfTTzPv73kX9Ryb4jeRmJ5Iw4oNaVixISaDyat7/ey7ZzOk6RDXF8icJm+YzDe7l+If4IOvr+mC7hMmk/GCz72USnwy9+VyMZviepK9nPxP+36kSZUmvLnqTa4Jvoa+jftyOv008SnxhAWEUSGwAv4mfxIsCaTaUrk25FoWbFvAQ4vyn3465545NKvSjJ5f9uSX/b8AWTeJTf028eaqN11d8GHmMPpH9ad/VH/XJpHhQeFUDKrIT3t/4qFFD3mckTIgagCjW4929W5VCa5CtdBqXtUv5yaS2VON77nuHnou7Om6ETx363Ms37OcHgt7eHW98FLhXF/heiyZFlJtqZQPLO9a7O1E6gn8TH6UNpcm057J1uNbubH8jVjsFrfNaZ04SbQkcjr9NGHmMMICwth9ajcVSlWg/WftSbWlsqzHMn498KvH3cBvq34b5QLLMeS7IYQFhuXaWNNoMHJjhRtz7fk0s8tMDiYd5LGbHuP/fv6/C96xHLI243xh5QusOrDK7Xhem5cCbO3/J1X8a2KzXZ49lq42Rbkprq+viQOWPTSa2jD/wv/5rc9vBPgG8P2/3/NIxCN8t+c7xqzOve1EziUCsrcUalixIV898BUHkw/mu1l3TqseXcX0P6bzQvMXuO59z9Nm89qs1tsNbIfFDOO6ctfR68te+dZnZpeZ7D69m8cbP35BG9PmNUT/WbfPqF2mNo8veTzPLaOaV23u9ru8kOfY2n8rTpz4mfzctnYBct3rywaWJTwonHl/zePFlS/mupdkB4pPRD2BEyen0/7bH84/a0cAR6YTgyHv7ZTc95Y7O8My+9xLuZuAZr1dZpciUIKzf0QmfycOp52lO5cy7Y9pdK3flftvuJ80WxrHUo9lBS2lKmJwGMm0OcEv0+OeYdkGRA1gWMwwdp/eTVSlqKz9yP4b+y8TUAYnTpbsXOJ24zMajNx73b08HPEwDSo0oM9XfZjScQprDq7hrbVv5d6Xp9lQOtbtSPcF3dl+cjvLeixj3eF1NKvSjDdWvZHvrt5vtnmTw8mHCfILIsQ/BLvTjslg4qt/vnIFWmUDyrLx8Y1eX+9g0kFC/EMo5VeKr/75ipmbZvLOXe+wbPcy0jLTOJx8mE3xm1w9b09EPcGYtmM4nXYaX6MvTpxMXj+ZJ6KfYOaWmZQLKEeHuh0I9g8m2ZpMlfFVgLyDy3N3A1/fdz2BvoGsOrCKmyvfnO9ryJ718vuR32lVrRWptlR+2veTx2Ds3A+nc+UVEEWGR/Ll/V9Sc2LNXDfYg88eJNAepr3jLpGiDJRMJiNppgSqvFPF63P+eeoffI2+NJ7WmL8G/IXdaceJ0+3LVLnAcoxZPYYpG6a4fcCaDCYODjrImYwz1HuvntfPuf3J7by56k0eu+kxjwFW9gy3cwOXMHMY6/uuJ25NnMf3mclg4vBzh7nz4zu9DnqWdl9KkjXpvIFbXq8jry8jA6IG8FKLl7hx0o0ev3zm9wUov+c497HsxPcW1Vq47hfnflGbd++8PK+Vcy/Qib9PdN2HwsxhvHrbq/Ro0IM0W1qu4MlmtWM0Zu8t93WR7C2nWW8lhNMJGRY7WMDHx0SnWnfTqW4nEi2JHE89QYXA8jSs0BC7zUBqUqYr8jZlGBjbdiwtq7fM8xtedgBjwECN0BpsOZY1ZbdqSFV2nt6Jv48/szfPJsmaxC+P/MLJtJMkWhIJ8A1wfdD3XNgza8huWhT9o/qz+IHFWDItpNnSKO1fGqPByNEzR0myJLmmxEZPj6Z/VH9aVW9FXJs4WlZrmed+TIOaDOK26rcxZf0UIsIjiKkWw1tr3uJY2jGGNhtKuzrteLzx40zbOI1T6ado83EblvVYxq3X3nreHpxRv4ziWOoxthzb4jYEGRYQxub4zXy+7fNc9WhZrSU9FvTgtOW0q2enUcVG9IvuR+XgyjSv2pxWs1sRdU0ULzR/wXV+dg/amNVj8uwZgqyFP/1MfrSu2Rpfoy9j246lVbVWed40hjQdQoc6HbDZbbSv3Z5GUxpxOv2063dvwECSNQk/kx/+Jn/39Wvy4Cl3ydNQXWR4JKXNoViSFSSVRHa7g9KlQgu0d5u/yR+DwUC10Gr0W9qPSe0nUWNCDdff+zVB1zDujnGMXjU69/M57UzaMIlBTQYV6DkDfALYcmzLeXMTs/Ofzg10soeXznevGNlyJBmZGQXenDfBi83Nc0q0JLoNb0eER/Dcrc9xV+27uHXGrecNgM6dnepJXkPokPu9nz2UmL1faHYAlnPj3rzuFZ72As0ZPLX+qLXHAMhgMDBkxfNFurecN9SjVEguVY+SJ97sDO/qjfpvXaacGwsG+JrxMfpw8uQZAoJNTNk4mY+2fuT6EM/+Q//1wK+kZKRQI7QGEeERfLjpQ0b+PDLPrlyjwcjIliPpHdmbl398mTtr3knrWq3ZeWonzT5slqts0ypN6dmgJx3rdiQlI4UzGWeoWKpi1jh82mkynZmUCSiD2ceMESPzt83noUUPEWYOY3jMcPo27svSnUtdQUV2D86T0U9ic9g4Yz2Dv4+/K2g49xtttsjwSH565CcAkq3JxKfEu2b/eTpvTrc5dK7XmQ82fcDIn0aSYEmgYcWGfH7v59R/v77X7Xjg2QMYDUYyHZlM2ziNYynHiLsjDgMGTqSd4ETqCcoGliXIL4i1B9cS7Bfs2ji02+fdXDeyxpUa8/Mjv+DMNPLr4R/pOKdjvrPucg5rZjvfN9Xs/Z0yLBp2u1SKskcJwM9s4pt9i+m+sHu+ZbP/fgwGA7XDatP3677sfWYvned2dn0wmgwmjj53lDs+uSPPQCjMHMaOp3awfPdyr4biZ989mx0ndxC3Oo5/n/k318bZ2ZpXbc7o1qM99jid29ubkpFCucByGDCwfM9ybih/A81nNs+3Ptn+ePwPgvyCqPue94tJHnj2AJmOTI6nHqdScCWcTidBfkF0ndeVXw/8mu/5+Q2tgecepbze++d7zFP5vIYyw8xhbHh8g8cNxLPlXMMvv4VEL9W9x9sepeKVMSVes9sd2Gz28w6BZPdGpSc7CbSHUcW/JoH2MCzJTtfMOKcTrKkOHr+pHy82e5GGFbPyE7KTipOtyfRo0IOYajFM3TiV7g26M6PzjDwT/T7t9ikDbxmI0WDknbveoVP9TgT6BnJNcO4Vvx1OB6sOrKL/0v5Ufacq//vifwxaPojOczuTlpFGsH8wlYIqYfYxs/vUbtp/1p4u9bvQv3F/EiwJPL/ieWpOqMnxtOMs7b6Uf576h68f/Jo+kX0I9g/mWMoxjqUc477597nN8MtLbLNYDiUfYtCyQTidTn7c9yP/++J/Hs8bEDWAjnU7Alm7cGc//vfxvwnwDShQYrnRYKTjnI7UnFiTN1a9wYzNM6g9sTbj1owj2C+Ya0OuJd2WTkpGCnfWvpN/Tv9D9PRo4lPi3b4pPt/0eTKtYEmz0bxKDI/fVLCZKzln1OWVqDkgagAd6nTEZlWQVJLZrHY61ulE/8aeE43B/e9n8vrJ3F7jdvre1JcpG6e4TTG3O+1M3TDV46y4BEsC7T89+94+nwFRA+havyuPRDzCz4/8jJ/Jj9imsXmWzW827LmzgV/58RXKBpRl9KrRvL/+fcoGlj1vXc4V4BvgdaI4ZL33HU4H3T7vxuQNk7E77Nw9725OpJ3wKkiC80/GyH6OvHqBzn3vn+vc2bADogZk9e6fUz57eYdzE+q9TWrPWqZg6XmT2rMV9Yxb9SgVksvdo3Sxzv3mem5CXc7eJ0eGERxgMjv47t/vOJR8iLvr3+3KQyhfqjxlA7JuLO/9/h6ZjkxMBhNNqjTh5mtuBgO5ul89ye7dSbelY8DAst3LaFWjFaN+HcXGIxtZ0n0JP+37Kc+8qUcjH+WG8jfQ5uM2nEg9kSshPS8DogbwQvMXmL15Ng9HPIzNbqNMQBmW7V6W5/DX0KZDswKGNHA4nPgHG4hdcTYXzNtkUchKpD+eepxnlj2T5+M5cwSuCbqGuLZxrryhnN8UB0QNYGzbsaQnO3A6z04CWLprCWPWeB52TbGmcCLtJGHmUELNoew6vYtnlz/rSurP6zVfijwBOauoe5TgAv9+AsLA6eS7Pd/R+JrGbrl2YeYw/uj3B2NWjfH4XmxcqTHLey5nxb8rPCaDt6reCiNGNsVvYvE/i9l1ehezuszKM98oO0fJU4/TuSLDI/mu53ckWZNYdWAVbWq2ObtKvhfnfnn/l0z/Yzq1wmp59d7P2UOTXdeRP2X1xhc0sT2vyRhwtvctuxfI27xFyOqJGvHjCNrVacddte9i9YHV3Dv/XrcyDSs2ZN6989zysjzlhnniTa9YtkuRH6lk7svsSg+UcvI0rHduMGXJtBDiH0KQX9bSqgmWBE6nJ7gCLGdG1jcSH39Y/I933flzus2hU+0uYHSyZOfXvLnmTVIyUlzDgDM3z6R5leY8Ef0EabY0EiwJVCxVkTIBZbA5bJxMO0lCeoIridThdLBk5xKPiYK3XHsLd35yp1vyYuXgytxZ+0661u9Kmi2NdFs65QPLu4LGnLMwzv1Q2Z+4P99kUcgK0OLaxLF8z/JcN6C85LyxZt+Yu33ejSG3DqFTvU6U8ivl1pb5Bb5ZiZRn29nhcORbXneKS684BEpQ8L8fu92R58ST7Pdd7TK1Wd5zOWsPrmXc2nEegy+b3YbJaCLZkpw1sSQgjNL+pVnx7woe++oxwH1GVmpGKrXL1Obb3d/mCrAmtptImDnMq5lrs7rMol2ddnz/7/ecTDtJjwY9vJ5JO+eeOdxW/TYOJB2gbtm6vPD9CwWeifpuu3epFVaLGmE1Ci0h/M02bxKfEk9CeoJraO98KQg5bem/hQqBFZj4+0Tm/jWXFb1WMG7NOLfXldfwZl7B04W+hnNdihm3CpQus5IUKHkjr2DqfAHWuT0veclO2rOmZP1J5rxRJ1vPEB5UET8fPxLTs27cFQLLE+wfnDXTz5BJkiWR9EwLpf1DKOUXRKYVDE7wNRuw2q2cTD/JqbRThAWEYfYxE+gbyOhfR/P2b2/nWR+jwcj/tfo/Bt0yCGuak8zM8+eCZdc11ZZGpeBwvtn1Td6J9Dl6aHwDPW8Xke3cG+tn3T7jzlp34mvyw5FhpHRw1h4dntrSm3y2iykvhae4BEo5Xcjfg4+PEV9/I5jsrvzIMgFZy2g4nA7X0hplzGUIDQjFx+BDsiUZf19/fE2+bDq6iQ83f8jx1OO83+59Av0C+WbXN4z/bTyb4ze7vtQ0rtSY/13/P2665iYy7BmkZKSQaEnM+qIUUA4MeDX7d2izodw2+zYeuPEB+kf1J8OeQfnA8l4HPY8sfoQ6ZerQ/cbu3HztzXy5/UveWfeO1zNRP7/3c9rWaovVbuWuT+7yuifrqwe/oto71Vy9MTmDzp4Le7I3cS+R4ZGMaDmC2u/W9mq3AMiazXj/F/e76tG5bmc+6fYJX+740tUGeQVF+eWG5eV8vWI5qUepBLjaAqWCyrc7/zzDO+feqD3duPPrCfPxh9SMFJKsyQT4mF3DBUt2FrxOnl9nVh2cTidGH0O+PTT5/V7OvbFm90TZ0gyuwK04frjKhSmJbZnX+zIw0A8fHyMOh5P0dBsmkxEfHyMGA2RmOjD5GHCaMkm0JnHGmkx4UDj+Pv7YHVmBV5I1idL+pQk1h7oWS0ywJFD+vy9PDqeDk2knCfINwuxrzrXMCWS9v56/9Xna1mqLj8GHVFuqa3mUsIAwfAw+ZDozPZ47tNlQ2tVux+7TuynlV8o1G3juX3N56463aFurLYmWRI6eOUqwf7DbsiA5e3T6Ne5HbLNYfI2+BPkFsXTXUq96wT7p+gmd63Um0ZLIoeRDrjX1Um2p7Dy1kwNJB+j9Ve8LGoY8d3mQOffMoVOtLhhMTuzOrDZItCRSPbQ6LWa1cF33UvUoRYZH8vMjv2BJLtxwRYHSZaZAKX/edOdf6r/Gc2/al6NO+X0jd5udaE3kROoJSvmVcputVy20msfArSR+uF6t1Jbust87BkPWlw9/fx9MJgNOpwGHw4HT6cThcOJ0gtFowGg04nC4f0lyOJwYTOA0ZpJoTSQxPZHQgKy8PCPG/1IGTlOhVAVC/ENwOB0YDUZ8Db4AOHCQ4chwBQel/UsTFhDGzlM7eW/de+xK2OVa9qNhxYY83/R5OtXtxMK/F9Llui74m/z5ef/PvPTjS/xx9A9X3bKDrfa127MvcR8VgipQyrcUdqedF79/Md9esDfbvMnE3yayfM9yMMAdNe/gsZsew5JpofVHrVnRa4UrBaAguZPnznDL2dOf9eUuK6gFMJhgyb+LXGkVhRGU5aWoZ70pUCokCpQKpjgO7xSHOmUNWZgw+DhISD/tVeBW1G0phUdteWnl/JDPzHS4fnY6cX1xMhqNQFYAlpnpcPViZ/d4GUyAjx2nM2u16YT0BLeeKJPBREJ6AifTTxJmDqNsYFlMBhMZ9rPBVlhAGKHmUFKsKRgwUMqvFA6yArQDiQeoUrqKx9zKwU0G06JaC+745A52ntrpWu+tfZ32pGWk0frj1uw+vdttiZeZm2fyYecP882dzDnMXy20GkOb/fflLDXvXnWDAfyDDG7pAwWd0OJv8qfb5908ljk3UCtMCpQuMwVKUti8DdzUliWH2vLKYTIZ8fPzcQ0h2u2O/4ItA76+RgIDszbgzW5Ls9n3v6DMicWSCWR9Mcp+j2dmOtwmVdgdmRiNRpItySRaEwn1/68XzHC2FyzUHEppc2l8DFnLvSzbvYzXfn3NbXXska1G0rNhTzLsGQT4BOSZ+O7q3arTnviUeAJ9Awn1z/pylpFPr/rFTGgZ0yYOnBeWklEYFChdZgqUpKioLUsOtWXJUVgTZrKHHc+dOJOzZ8ybNAJHpjNreNLHAD72XGWMDh+cdtx60rx1MRNaHA5nkaVkKFC6zBQoSVFRW5YcasuSo6j37cuvN/pSpRoUdELL5aiTJ9rrTURE5CqUNQx48WUu+rkzAQsEmsII8i+Dw+78b6/IvJ/4UtXpYilQEhERkUumuAZA3tJebyIiIiIeKFASERER8UCBkoiIiIgHCpREREREPFCgJCIiIuKBAiURERERDxQoiYiIiHiglbkLidOZtez7lSJ7NVKb7Qpe3EIAtWVJorYsOdSWxV/WRseGfMspUBIRERHxQENvIiIiIh4oUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIiIi4oECJREREREPFCiJiIiIeKBASURERMQDBUoiIiIiHihQEhEREfFAgZKIiIiIBwqURERERDxQoCQiIiLigQIlEREREQ8UKImIiIh4oEBJRERExAMFSiIiIiIeKFC6yhw9epTBgwfTrFkzoqOj6dOnD7t27XIrs3btWrp160bDhg254447WLRoUdFUVvLlcDiYOHEiMTExNGrUiN69e7N///6irpbkIzExkVdeeYUWLVpw00038eCDD7JhwwbX49u3b6dnz55ERETQqlUrZsyYUYS1FW/t3buXyMhIFi5c6DqmtrzyKVC6imRkZPD4449z6tQppk6dymeffUZwcDAPP/wwp0+fBmDPnj3069ePli1bsmjRIu6//36GDRvG2rVri7j2kpdJkyYxd+5cXn/9debNm4fBYKBv375kZGQUddXkPAYPHsyWLVt4++23+eKLL7jhhhvo06cPe/bsISEhgUcffZTq1auzYMECnn76aSZMmMCCBQuKutpyHjabjeeff560tDTXMbVlCeGUq8bq1auddevWdcbHx7uOWa1WZ6NGjZzz5893Op1O58svv+y877773M4bPHiws3fv3pe1rpI/q9XqjIyMdH722WeuY0lJSc6GDRs6lyxZUoQ1k/PZt2+fs27dus6NGze6jjkcDmfbtm2d77zzjnPKlCnOmJgYp81mcz3+1ltvOe+8886iqK546a233nL26tXLWbduXeeCBQucTqdTbVlCqEfpKlKnTh2mTZtGxYoV3Y47nU6SkpIA2LBhA02aNHF7vEmTJmzcuBGn03nZ6ir527FjB6mpqW7tFRISwvXXX8/69euLsGZyPmFhYUybNo0bb7zRdcxgMLjehxs2bCA6OhofHx/X402aNGHv3r2cOnWqKKos+Vi/fj3z5s1jzJgxbsfVliWDAqWrSPny5WnZsqXbsY8++gir1UqzZs0AiI+PJzw83K1MhQoVSE9PJyEh4bLVVfIXHx8PQKVKldyOV6hQgaNHjxZFlcQLISEhtGzZEj8/P9exb7/9lgMHDtC8eXOP70GAI0eOXNa6Sv6Sk5OJjY3lpZdeyvVeVFuWDD75F5ErxaFDh2jdurXHx1etWkX58uVdP3/33XeMHz+eXr16Ub9+fQAsFovbDRxw/ay8l+IlPT0dIFd7+fv7u3oIpfjbuHEjw4YNo3Xr1tx+++2MHj06zzYFsFqtRVFFOY+RI0cSERFBp06dcj2W1/1UbXnlUaBUglSsWJFvvvnG4+NlypRx/f+cOXN47bXXaN++PS+++KLruL+/f66AKPvngICAQq6xXAyz2QxktU/2/0PWDVhtdWX4/vvvef7552nUqBFvv/02kNWu574Hsz9UAwMDL3sdxbNFixaxYcMGvv766zwfV1uWDAqUShBfX19q1aqVb7lx48Yxffp0evXqxfDhwzEYDK7HKlWqxPHjx93KHz9+nMDAQIKDgwu9znLhsrv5jx8/TtWqVV3Hjx8/7uohlOLrk08+YdSoUbRt25Zx48a5eh7Cw8PzfA8CufILpWgtWLCAU6dO0apVK7fjI0aMYMaMGVxzzTVqyxJAgdJVZuzYsXzwwQfExsbSp0+fXI9HRUXx+++/ux1bu3YtN910E0ajUtqKk/r16xMUFMS6detcgVJycjLbtm2jZ8+eRVw7OZ/PPvuM1157jV69ejFs2DC391Z0dDRz587FbrdjMpmArPdgjRo1KFu2bFFVWfIwbtw4LBaL27E77riDgQMH0r59e5YuXaq2LAH0yXcVWbduHR988AG9evWic+fOnDhxwvUvNTUVgF69erF161bGjRvHnj17+PDDD1m+fDmPPfZYEddezuXn50fPnj0ZN24cK1euZMeOHQwaNIjw8HDatm1b1NUTD/bu3csbb7xB27Zt6devH6dOnXK9D8+cOcM999xDSkoKw4cPZ/fu3SxcuJDZs2fTr1+/oq66nKNixYpUq1bN7R9A2bJlqVy5stqyhDA4Nef7qvHyyy/z+eef5/nYU089xdNPPw3AL7/8wtixY9m3bx/XXnstTz/9NO3bt7+cVRUv2e123n77bRYuXIjFYiE6OppXXnmFa6+9tqirJh5MmTKF8ePH5/lY165defPNN9m6dSujRo1i27ZtlC9fnt69e6uX8ApRr149Ro8eTbdu3QDUliWAAiURERERDzT0JiIiIuKBAiURERERDxQoiYiIiHigQElERETEAwVKIiIiIh4oUBIRERHxQIGSiIiIiAcKlEREREQ8UKAkIhfl9ttvp169eq5/1113HVFRUfTq1YsNGzYU+vOtW7eOevXqcejQISBr250XXnjBq3PT0tL49NNPL+r5Dx06RL169Vi3bt1FXceTc1+fiBQtbYorIhetd+/e9O7dGwCn00liYiJvv/02jz32GMuWLSM8PPySPfe7777r2nA0Px9++CELFy6kR48el6w+IlKyqEdJRC5aYGAg5cuXp3z58lSoUIG6devy6quvkp6eznfffXdJnzs0NJTg4GCvymrHJhEpKAVKInJJ+PhkdVj7+fkBWUN0b7zxBu3bt+eWW27ht99+w+l0Mn36dFq3bk2jRo3o0qULX331ldt1NmzYwH333UfDhg25++67+eeff9weP3fo7a+//uLRRx8lMjKSpk2b8sorr5CWlsa7777Le++9x+HDh92GthYsWEC7du1o2LAh7dq1Y/bs2TgcDtf1du7cyUMPPURERAR33nknv/32m8fXnJqaSmRkJJ999pnb8cmTJ9OqVSscDgfJycmMGDGCli1bcsMNN9CsWTNGjBiBxWLJ85p5DS2+8MIL9OrVy/XzsWPHGDRoEFFRUdxyyy3079+fffv2uR4/deoUAwcO5JZbbqFhw4Y88MAD/P777x5fh4icpaE3ESl0x44d44033iAwMJAWLVq4js+ZM4epU6cSHBxMvXr1GD9+PF9//TWvvPIKtWrVYv369YwcOZIzZ87Qo0cPDh48SO/evbn77rt588032b17N6+88orH5z106BC9evXi9ttvZ968eaSkpPDiiy/yyiuv8Oqrr5KWlsY333zDF198QZkyZZg3bx5vvfUWr7zyCo0aNWLbtm289tprHDt2jNjYWM6cOcMjjzxCREQE8+fP5/jx47z88ssen79UqVLceeedfP3113Tv3t11/Ouvv6ZLly4YjUaGDh1KfHw8EydOpGzZsmzevJkXX3yRmjVr8vDDDxf4d52WlkavXr2oX78+n3zyCUajkZkzZ/K///2Pr7/+mooVKzJy5EisViuffPIJfn5+TJkyhSeeeIJffvmFwMDAAj+nyNVEgZKIXLSpU6fy4YcfApCZmUlGRga1atXinXfe4ZprrnGVa9myJU2bNgWyPuBnzZpFXFwct912GwBVq1bl8OHDzJgxgx49evD5559Trlw5RowYgclkolatWhw9epTRo0fnWY/PP/+c0qVL8+abb+Lr6wvA66+/zu+//06pUqUIDAzEZDJRvnx5ACZNmkS/fv3o2LEjAFWqVCElJYVXX32VZ555hqVLl5Kens6YMWMIDg6mTp06DBs2jCeffNLj76Jbt2489NBDHDp0iGuvvZa//vqLPXv2MGnSJACaNWtGVFQU9evXB+Daa6/lk08+ydVT5q2lS5eSkJDAW2+95XrNo0aNYt26dXz++ec8/fTT/H879xPS9B/HcfzpmLOC4WHGCEtNUrBOiqtDUSczFw4nHdTTCAWFhVJZrAgJo4SKtlF0qaCsDh0qKeqwWA3BQ3UwRiNHY/CdiFAYYWGzP7/fIfzSqpX9focOvh4w+PL5/vl8Prvsxefz/s4wDKqrqykrK6OoqIjDhw/T3Ny86NoukaVMQUlE/re2tjZzK8hiseStGyovLzePX758STab5eDBgwQCAbN9IWh9+PCBZDLJ+vXrc37Q6+rq8o5jYmKCDRs2mIEBwOVy4XK5frh2ZmaG6elpQqEQZ8+eNdu/fPlCNptlcnKSZDJJRUVFzlxqa2t/+V24XC5Wr17N3bt36e7uZmRkhNraWioqKgDo6OggGo0yMjKCYRgkk0kymYx5/k8lEgnevXvHxo0bc9qz2SypVAoAv99Pf38/kUiE+vp6tmzZgtvtpqio6D/1KbKUKCiJyP9WXFycE4LyWbZsmXm8UFgdDAaprKz84dqF2qbvC7AXap9+xmq1UlBQsKgxL9QhBQIBc5XrW6tWrfrj/gEKCgpoaWnhzp07dHV1ce/ePXp7e81ndXd3MzExQXNzM42Njezdu/eX23k/G8PHjx9z5rF27VrOnz//w30L22oNDQ2Mjo4yOjrK2NgYFy5cIBQKcePGDaqqqn7Zt8hSp2JuEfkrKisrsVqtTE1NUV5ebn5isRgXL17EYrFQU1NDPB5nfn7evC8ej+d95rp160gkEnz+/Nlsi0QibN26lbm5uZwQ5XA4cDgcGIaR0//z588JBoMA1NTUkE6nmZmZWVT/C7xeL6lUiuvXrzM7O0tTUxPwdfUnFosRDofZv38/Ho+HsrIyDMPI+0ZeYWEhs7OzOW2GYZjH1dXVTE1NYbfbzTmUlpZy+vRpnjx5wvz8PCdOnCCTyeB2uzl27BiRSASLxcKjR49+OxeRpU5BSUT+CrvdTltbG8FgkNu3b5PJZLh16xYnT56kpKQEgPb2dubm5jh06BCpVIqHDx/mbJN9r6Ojgzdv3jAwMEAqleLp06ecOnWKzZs3s3z5clasWMHbt29Jp9N8+vSJzs5OhoeHGR4exjAMHjx4wNGjR7HZbNhsNnbu3InD4WDfvn28ePGCx48fc/z48d/OrbS0lE2bNnHmzBkaGhrMrbuSkhKsViv3798nk8kQj8fp6+vj1atXOWHwW3V1dYyNjRGNRslkMoTDYZLJpHne4/FQXFyM3+9nfHycVCpFIBAgFotRVVWFzWbj2bNnHDlyhPHxcSYnJ7l586b5hp6I/JqCkoj8NYFAAJ/PRzgcpqmpiXPnzuH3+9mzZw8ATqeTy5cvMz09jdfrZWhoiJ6enrzPczqdXLp0iXQ6jdfrpa+vj23btjEwMADA9u3bWblyJR6Ph0Qiwe7duwkEAly7dg23283g4CCtra0MDg4CX7eurly5QmFhIe3t7Rw4cICurq5Fza21tZX379/T0tKSM76hoSGi0Shut5ve3l6cTic+n494PP7TVSWfz0djYyP9/f14vV5ev36Nz+czz9vtdq5evYrD4aCzs5Ndu3aZBfEL22qhUIg1a9bQ09PDjh07zLf96uvrFzUXkaWs4B/9A5uIiIjIT2lFSURERCQPBSURERGRPBSURERERPJQUBIRERHJQ0FJREREJA8FJREREZE8FJRERERE8lBQEhEREclDQUlEREQkDwUlERERkTwUlERERETy+BcKWP0YBMczewAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the residuals against the predicted values\n",
    "\n",
    "residuals = y_test - y_pred_lm\n",
    "\n",
    "sns.set_theme(style='darkgrid', context='talk')\n",
    "sns.scatterplot(x=y_pred_lm, y=residuals, color='green')\n",
    "plt.axhline(y=0, color='blue', linestyle='--')\n",
    "plt.xlabel('Predicted values', fontsize=11)\n",
    "plt.xticks(fontsize=11)\n",
    "plt.ylabel('Residuals', fontsize=11)\n",
    "plt.yticks(fontsize=11)\n",
    "plt.title('Residual Plot', fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This model gave us a very low Mean Absolue Error (MAE) of 18.87. This means that, on average, the model's predictions are off by around 18.87 minutes from the actual arrival delay times. This might seem a large error, but at the same time this result is going to depend of the context we are evaluating our problem. In second place, we got a Mean Squared Error (MSE) of 3218.57. This means that, on average, the squared difference between the model's predictions and the actual arrival delay times is very large. As this metric is very sensitive and influenced by outliers, we would need to investigate whether the model is underestimating or overestimating arrival delays at another moment. In third place, we got a very low R-Adjusted Square of 0.16. This means that around 16% of the variance in the target variable (arrival delay times) is explained by the model. This is a relatively low R2 score, indicating that the model is not able to capture a significant portion of the variability in the data. Finally, in regards to the Root Mean Squared Error (RMSE) of 56.73, as this indicates a medium/high result, we should consider another models to see if we can find better performance as the MSE provides an estimate of the standard deviation of the residuals (prediction errors). "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's try now other models to see if they offer a better optimization and prediction to our flight delays."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 782,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Ridge Regression Model Results are:\n",
      "Mean Absolute Error (MAE) is 18.87.\n",
      "Mean Squared Error (MSE) is 3218.56.\n",
      "Root Mean Squared Error (RMSE) is 56.73.\n",
      "R2_score is 0.16.\n",
      "Best alpha: 6.0\n"
     ]
    }
   ],
   "source": [
    "# Checking the best alpha parameter value with Ridge Cross Validation\n",
    "# After we are modelling with Ridge Regression\n",
    "\n",
    "from sklearn.linear_model import RidgeCV\n",
    "\n",
    "ridge = RidgeCV(alphas=[0.1, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0, 4.5, 5.0, 5.5, 6.0])\n",
    "ridge.fit(X_pca_train, y_train)\n",
    "y_pred_ridge_cv = ridge.predict(X_pca_test)\n",
    "\n",
    "# Calculating some evaluation metrics\n",
    "\n",
    "mae_ridge_cv = mean_absolute_error(y_test, y_pred_ridge_cv)\n",
    "mse_ridge_cv = mean_squared_error(y_test, y_pred_ridge_cv)\n",
    "rmse_ridge_cv = np.sqrt(mse_ridge_cv)\n",
    "r2_ridge_cv = r2_score(y_test, y_pred_ridge_cv)\n",
    "\n",
    "print('The Ridge Regression Model Results are:')\n",
    "print('Mean Absolute Error (MAE) is {:.2f}.'.format(mae_ridge_cv))\n",
    "print('Mean Squared Error (MSE) is {:.2f}.'.format(mse_ridge_cv))\n",
    "print('Root Mean Squared Error (RMSE) is {:.2f}.'.format(rmse_ridge_cv))\n",
    "print('R2_score is {:.2f}.'.format(r2_ridge_cv))\n",
    "print(\"Best alpha:\", ridge.alpha_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 783,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Ridge Regression Results are:\n",
      "Mean Absolute Error (MAE) is 18.87.\n",
      "Mean Squared Error (MSE) is 3218.56.\n",
      "Root Mean Squared Error (RMSE) is 56.73.\n",
      "R-squared is 0.16.\n"
     ]
    }
   ],
   "source": [
    "# Modelling with Ridge Regression using an alpha value of 0.1\n",
    "\n",
    "from sklearn.linear_model import Ridge\n",
    "\n",
    "ridge = Ridge(alpha=6.0) # alpha is a parameter to prevent overfitting\n",
    "ridge.fit(X_pca_train, y_train)\n",
    "y_pred_ridge = ridge.predict(X_pca_test)\n",
    "\n",
    "# Calculating some evaluation metrics\n",
    "\n",
    "mae_ridge = mean_absolute_error(y_test, y_pred_ridge)\n",
    "mse_ridge = mean_squared_error(y_test, y_pred_ridge)\n",
    "rmse_ridge = np.sqrt(mse_ridge)\n",
    "r2_ridge = r2_score(y_test, y_pred_ridge)\n",
    "\n",
    "print('The Ridge Regression Results are:')\n",
    "print('Mean Absolute Error (MAE) is {:.2f}.'.format(mae_ridge))\n",
    "print('Mean Squared Error (MSE) is {:.2f}.'.format(mse_ridge))\n",
    "print('Root Mean Squared Error (RMSE) is {:.2f}.'.format(rmse_ridge))\n",
    "print('R-squared is {:.2f}.'.format(r2_ridge))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 784,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1395.000000\n",
       "mean        4.974663\n",
       "std        24.648541\n",
       "min       -24.127155\n",
       "25%       -15.051752\n",
       "50%       -10.315125\n",
       "75%        34.319767\n",
       "max        54.265858\n",
       "dtype: float64"
      ]
     },
     "execution_count": 784,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Statistics\n",
    "\n",
    "pd.Series(y_pred_ridge).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 785,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the Ridge Regression Model\n",
    "\n",
    "sns.set_theme(style='darkgrid', context='talk')\n",
    "sns.scatterplot(x=y_pred_ridge, y=residuals, color='green', alpha=0.5)\n",
    "plt.axhline(y=0, color='blue', linestyle='--', linewidth=1)\n",
    "sns.regplot(x=y_pred_ridge, y=residuals, scatter_kws={'color': 'green', 'alpha': 0.5}, line_kws={'color': 'indigo', 'linewidth': 2})\n",
    "plt.xlabel('Predicted values', fontsize=11)\n",
    "plt.xticks(fontsize=11)\n",
    "plt.ylabel('Residuals', fontsize=11)\n",
    "plt.yticks(fontsize=11)\n",
    "plt.title('Residual Plot', fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Again, with this model from Ridge Regression, we found the results very similar to our Linear Regression Model. So we are going to save this model to the disk before moving forward to check other models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 788,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Saving this trained model to disk using pickle\n",
    "\n",
    "with open('ridge_model.pkl', 'wb') as f:\n",
    "    pickle.dump(ridge, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 789,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Loading the trained model from disk using pickle\n",
    "\n",
    "with open('ridge_model.pkl', 'rb') as f:\n",
    "    ridge_model = pickle.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Decision Tree Regressor and Random Forest Regressor Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 790,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Decision Tree Regressor Results are:\n",
      "Mean Absolute Error (MAE) is 26.48.\n",
      "Mean Squared Error (MSE) is 7120.65.\n",
      "Root Mean Squared Error (RMSE) is 84.38.\n",
      "R-squared is -0.86.\n"
     ]
    }
   ],
   "source": [
    "# Modeling with the Decision Tree Regressor\n",
    "\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "\n",
    "dt = DecisionTreeRegressor(random_state=100)\n",
    "dt.fit(X_pca_train, y_train)\n",
    "y_pred_dt = dt.predict(X_pca_test)\n",
    "\n",
    "# Calculating some evaluation metrics \n",
    "\n",
    "mae_dt = mean_absolute_error(y_test, y_pred_dt)\n",
    "mse_dt = mean_squared_error(y_test, y_pred_dt)\n",
    "r2_dt = r2_score(y_test, y_pred_dt)\n",
    "rmse_dt = np.sqrt(mse_dt)\n",
    "\n",
    "print('The Decision Tree Regressor Results are:')\n",
    "print('Mean Absolute Error (MAE) is {:.2f}.'.format(mae_dt))\n",
    "print('Mean Squared Error (MSE) is {:.2f}.'.format(mse_dt))\n",
    "print('Root Mean Squared Error (RMSE) is {:.2f}.'.format(rmse_dt))\n",
    "print('R-squared is {:.2f}.'.format(r2_dt))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Based on the metrics of the decision tree, it looks like the decision tree regressor is performing even worse the linear regression model in terms of MSE and MAE. The MSE and RMSE values are higher for the decision tree regressor, which means that the model's predictions are more spread out from the actual values and the linear model is still offering a better performance than this one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 791,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Random Forest Regressor Results are:\n",
      "Mean Absolute Error (MAE) is 26.48.\n",
      "Mean Squared Error (MSE) is 7120.65.\n",
      "Root Mean Squared Error (RMSE) is 84.38.\n",
      "R-squared (R2) is -0.86.\n"
     ]
    }
   ],
   "source": [
    "# Modelling with a Random Forest Regressor\n",
    "\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "rf = RandomForestRegressor(n_estimators=100, random_state=100)\n",
    "rf.fit(X_pca_train, y_train)\n",
    "y_pred_rf = rf.predict(X_pca_test)\n",
    "\n",
    "# Calculating some evaluation metrics \n",
    "\n",
    "mae_rf = mean_absolute_error(y_test, y_pred_rf)\n",
    "mse_rf = mean_squared_error(y_test, y_pred_rf)\n",
    "r2_rf = r2_score(y_test, y_pred_rf)\n",
    "rmse_rf = np.sqrt(mse_rf)\n",
    "\n",
    "print('The Random Forest Regressor Results are:')\n",
    "print('Mean Absolute Error (MAE) is {:.2f}.'.format(mae_dt))\n",
    "print('Mean Squared Error (MSE) is {:.2f}.'.format(mse_dt))\n",
    "print('Root Mean Squared Error (RMSE) is {:.2f}.'.format(rmse_dt))\n",
    "print('R-squared (R2) is {:.2f}.'.format(r2_dt))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Based on the metrics of the Random Forest Regressor, we got the same results of the decision tree which makes sense as the random forest is an aggregation of a group of decision trees."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Support Vector Machines (SVMs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 797,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Support Vector Machines Results are:\n",
      "Mean Absolute Error (MAE) is 15.93.\n",
      "Mean Squared Error (MSE) is 3404.39.\n",
      "Root Mean Squared Error (RMSE) is 58.35.\n",
      "R-squared (R2) is 0.11.\n"
     ]
    }
   ],
   "source": [
    "# Modelling with SVM with a linear kernel\n",
    "\n",
    "from sklearn.svm import SVR\n",
    "\n",
    "svm = SVR(kernel='linear')\n",
    "svm.fit(X_pca_train, y_train)\n",
    "y_pred_linear = svm.predict(X_pca_test)\n",
    "\n",
    "# Calculating some evaluation metrics\n",
    "\n",
    "mae_svm = mean_absolute_error(y_test, y_pred_linear)\n",
    "mse_svm = mean_squared_error(y_test, y_pred_linear)\n",
    "rmse_svm = np.sqrt(mse_svm)\n",
    "r2_svm = r2_score(y_test, y_pred_linear)\n",
    "\n",
    "print('The Support Vector Machines Results are:')\n",
    "print('Mean Absolute Error (MAE) is {:.2f}.'.format(mae_svm))\n",
    "print('Mean Squared Error (MSE) is {:.2f}.'.format(mse_svm))\n",
    "print('Root Mean Squared Error (RMSE) is {:.2f}.'.format(rmse_svm))\n",
    "print('R-squared (R2) is {:.2f}.'.format(r2_svm))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 798,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1395.000000\n",
       "mean       -2.063689\n",
       "std        13.370198\n",
       "min       -16.555639\n",
       "25%       -12.495349\n",
       "50%       -11.222061\n",
       "75%        14.778863\n",
       "max        18.561755\n",
       "dtype: float64"
      ]
     },
     "execution_count": 798,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Statistics\n",
    "\n",
    "pd.Series(y_pred_linear).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 799,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Creating a scatter plot of predicted values versus true values\n",
    "\n",
    "sns.regplot(x=y_test, y=y_pred_linear, scatter_kws={'alpha':0.3})\n",
    "plt.xlabel('True Values', fontsize=11)\n",
    "plt.xticks(fontsize=11)\n",
    "plt.ylabel('Predictions', fontsize=11)\n",
    "plt.yticks(fontsize=11)\n",
    "plt.title('True vs. Predicted Values (Linear SVM)', fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that the SVM is also not offering a got fit for our predictions. It looks like its performance is similar to the one offered by the Ridge Regression Model, but with a higher MSE, and a lower R-squared value (value of 0.11). However, in an overwall, the values looked very similar to the linear model and the MAE is a little bit lower than the linear model (it is 15.93) so its predictions are off only 15.93 minutes from the actual arrival delay times. Therefore we are going to save this model to our disk. One possible reason for the poor performance could be the choice of kernel. So we are going to try experimenting with different kernels, such as the polynomial to see if it improves the performance of the SVM model afterwards."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 801,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['svm_linear_model.pkl']"
      ]
     },
     "execution_count": 801,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Saving the trained SVM model to disk\n",
    "\n",
    "import joblib\n",
    "\n",
    "joblib.dump(svm, 'svm_linear_model.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 802,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Loading the saved SVM model from disk\n",
    "\n",
    "loaded_svm = joblib.load('svm_linear_model.pkl')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's try the SVM with the polynomial kernel."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 803,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Support Vector Machines Results are:\n",
      "Mean Absolute Error (MAE) is 16.52\n",
      "Mean Squared Error (MSE) is 3544.7\n",
      "Root Mean Squared Error (RMSE) is 56.73\n",
      "R-squared (R2) is 0.07\n"
     ]
    }
   ],
   "source": [
    "# Modelling with SVM with a polynomial kernel\n",
    "\n",
    "svm_poly = SVR(kernel='poly', degree=3, C=1.0, gamma='scale')\n",
    "svm_poly.fit(X_pca_train, y_train)\n",
    "y_pred_poly = svm_poly.predict(X_pca_test)\n",
    "\n",
    "# Calculating some  evaluation metrics\n",
    "\n",
    "mae_svm_poly = mean_absolute_error(y_test, y_pred_poly)\n",
    "mse_svm_poly = mean_squared_error(y_test, y_pred_poly)\n",
    "rmse_svm_poly = np.sqrt(mse)\n",
    "r2_svm_poly = r2_score(y_test, y_pred_poly)\n",
    "\n",
    "print('The Support Vector Machines Results are:')\n",
    "print('Mean Absolute Error (MAE) is', round(mae_svm_poly, 2))\n",
    "print('Mean Squared Error (MSE) is', round(mse_svm_poly, 2))\n",
    "print('Root Mean Squared Error (RMSE) is', round(rmse_svm_poly, 2))\n",
    "print('R-squared (R2) is', round(r2_svm_poly, 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 804,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Creating a scatter plot of predicted values versus true values\n",
    "\n",
    "sns.regplot(x=y_test, y=y_pred_poly, scatter_kws={'alpha':0.3})\n",
    "plt.xlabel('True Values', fontsize=11)\n",
    "plt.xticks(fontsize=11)\n",
    "plt.ylabel('Predictions', fontsize=11)\n",
    "plt.yticks(fontsize=11)\n",
    "plt.title('True vs. Predicted Values (Polynomial SVM)', fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With those results, the SVM with the polynomial kernel did not offer a better performance to our model. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### XGBOOST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 805,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: xgboost in /opt/anaconda3/envs/base_env/lib/python3.8/site-packages (1.7.4)\n",
      "Requirement already satisfied: scipy in /opt/anaconda3/envs/base_env/lib/python3.8/site-packages (from xgboost) (1.9.3)\n",
      "Requirement already satisfied: numpy in /opt/anaconda3/envs/base_env/lib/python3.8/site-packages (from xgboost) (1.23.5)\n"
     ]
    }
   ],
   "source": [
    "!pip install xgboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 806,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The XGBoost Model Results are:\n",
      "Mean Absolute Error (MAE) is 19.70.\n",
      "Mean Squared Error (MSE) is 3474.77.\n",
      "Root Mean Squared Error (RMSE) is 58.95.\n",
      "R2_score is 0.09.\n"
     ]
    }
   ],
   "source": [
    "# Modelling with XGBoost Regressor Model\n",
    "import xgboost as xgb\n",
    "\n",
    "# Creating a DMatrix for XGBoost\n",
    "\n",
    "dtrain = xgb.DMatrix(X_pca_train, label=y_train)\n",
    "params = {\n",
    "    'objective': 'reg:squarederror',\n",
    "    'eval_metric': 'rmse',\n",
    "    'learning_rate': 0.1,\n",
    "    'max_depth': 5,\n",
    "    'subsample': 0.7,\n",
    "    'colsample_bytree': 0.7,\n",
    "    'seed': 42\n",
    "}\n",
    "\n",
    "# Trainning XGBoost model and predicting on test data\n",
    "\n",
    "num_rounds = 100\n",
    "xgb_model = xgb.train(params, dtrain, num_rounds)\n",
    "dtest = xgb.DMatrix(X_pca_test)\n",
    "y_pred_xgb = xgb_model.predict(dtest)\n",
    "\n",
    "# Calculating evaluation metrics\n",
    "\n",
    "mae_xgb = mean_absolute_error(y_test, y_pred_xgb)\n",
    "mse_xgb = mean_squared_error(y_test, y_pred_xgb)\n",
    "rmse_xgb = np.sqrt(mse_xgb)\n",
    "r2_xgb = r2_score(y_test, y_pred_xgb)\n",
    "\n",
    "print('The XGBoost Model Results are:')\n",
    "print('Mean Absolute Error (MAE) is {:.2f}.'.format(mae_xgb))\n",
    "print('Mean Squared Error (MSE) is {:.2f}.'.format(mse_xgb))\n",
    "print('Root Mean Squared Error (RMSE) is {:.2f}.'.format(rmse_xgb))\n",
    "print('R2_score is {:.2f}.'.format(r2_xgb))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From these metrics, we can conclude that the Linear Regression Model or the Ridge Regression Model provides the best fit for our predictions for now; and, in the second place, the Support Vector Machine with the linear kernell, but in the future it would be needed new insights and approachs to our feature engineering as this step can help us to improve the performance of our models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You have data from 2018 and 2019 to develop models. Use different evaluation metrics for each problem and compare the performance of different models.\n",
    "\n",
    "You are required to predict delays on **out of sample** data from **first 7 days (1st-7th) of January 2020** and to share the file with LighthouseLabs. Sample submission can be found in the file **_sample_submission.csv_**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 570,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>predicted_flights_delay</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32.728111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>35.722451</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-1.152322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40.789254</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-16.436766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1390</th>\n",
       "      <td>-17.454642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1391</th>\n",
       "      <td>40.145060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1392</th>\n",
       "      <td>35.382037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1393</th>\n",
       "      <td>31.377493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1394</th>\n",
       "      <td>-13.829567</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1395 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      predicted_flights_delay\n",
       "0                   32.728111\n",
       "1                   35.722451\n",
       "2                   -1.152322\n",
       "3                   40.789254\n",
       "4                  -16.436766\n",
       "...                       ...\n",
       "1390               -17.454642\n",
       "1391                40.145060\n",
       "1392                35.382037\n",
       "1393                31.377493\n",
       "1394               -13.829567\n",
       "\n",
       "[1395 rows x 1 columns]"
      ]
     },
     "execution_count": 570,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Converting the predictions given by the Ridge Model to a df\n",
    "\n",
    "y_pred_ridge_df = pd.DataFrame(y_pred_ridge).rename(columns={0: 'predicted_flights_delay'})\n",
    "y_pred_ridge_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 571,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin_airport_id</th>\n",
       "      <th>origin</th>\n",
       "      <th>dest_airport_id</th>\n",
       "      <th>dest</th>\n",
       "      <th>...</th>\n",
       "      <th>quarter_Q3</th>\n",
       "      <th>quarter_Q4</th>\n",
       "      <th>dep_time_of_day_afternoon</th>\n",
       "      <th>dep_time_of_day_evening</th>\n",
       "      <th>dep_time_of_day_morning</th>\n",
       "      <th>dep_time_of_day_night</th>\n",
       "      <th>arr_time_of_day_afternoon</th>\n",
       "      <th>arr_time_of_day_evening</th>\n",
       "      <th>arr_time_of_day_morning</th>\n",
       "      <th>arr_time_of_day_night</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3153</td>\n",
       "      <td>19</td>\n",
       "      <td>2100</td>\n",
       "      <td>3153</td>\n",
       "      <td>15096</td>\n",
       "      <td>SYR</td>\n",
       "      <td>14100</td>\n",
       "      <td>PHL</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>114</td>\n",
       "      <td>4</td>\n",
       "      <td>1997</td>\n",
       "      <td>114</td>\n",
       "      <td>13796</td>\n",
       "      <td>OAK</td>\n",
       "      <td>12478</td>\n",
       "      <td>JFK</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>3276</td>\n",
       "      <td>18</td>\n",
       "      <td>3831</td>\n",
       "      <td>3276</td>\n",
       "      <td>11027</td>\n",
       "      <td>CKB</td>\n",
       "      <td>12264</td>\n",
       "      <td>IAD</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>283</td>\n",
       "      <td>23</td>\n",
       "      <td>3806</td>\n",
       "      <td>283</td>\n",
       "      <td>10821</td>\n",
       "      <td>BWI</td>\n",
       "      <td>11193</td>\n",
       "      <td>CVG</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1021</td>\n",
       "      <td>23</td>\n",
       "      <td>3001</td>\n",
       "      <td>1021</td>\n",
       "      <td>12339</td>\n",
       "      <td>IND</td>\n",
       "      <td>11292</td>\n",
       "      <td>DEN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
       "0                   0            0                3153                 19   \n",
       "1                   2            2                 114                  4   \n",
       "2                   8            8                3276                 18   \n",
       "3                  10           10                 283                 23   \n",
       "4                  10           10                1021                 23   \n",
       "\n",
       "   tail_num  op_carrier_fl_num  origin_airport_id origin  dest_airport_id  \\\n",
       "0      2100               3153              15096    SYR            14100   \n",
       "1      1997                114              13796    OAK            12478   \n",
       "2      3831               3276              11027    CKB            12264   \n",
       "3      3806                283              10821    BWI            11193   \n",
       "4      3001               1021              12339    IND            11292   \n",
       "\n",
       "  dest  ...  quarter_Q3  quarter_Q4  dep_time_of_day_afternoon  \\\n",
       "0  PHL  ...           0           1                          1   \n",
       "1  JFK  ...           0           1                          0   \n",
       "2  IAD  ...           1           0                          0   \n",
       "3  CVG  ...           0           0                          0   \n",
       "4  DEN  ...           0           0                          0   \n",
       "\n",
       "   dep_time_of_day_evening  dep_time_of_day_morning  dep_time_of_day_night  \\\n",
       "0                        0                        0                      0   \n",
       "1                        1                        0                      0   \n",
       "2                        0                        1                      0   \n",
       "3                        0                        1                      0   \n",
       "4                        0                        1                      0   \n",
       "\n",
       "   arr_time_of_day_afternoon  arr_time_of_day_evening  \\\n",
       "0                          1                        0   \n",
       "1                          0                        0   \n",
       "2                          0                        0   \n",
       "3                          0                        0   \n",
       "4                          0                        0   \n",
       "\n",
       "   arr_time_of_day_morning  arr_time_of_day_night  \n",
       "0                        0                      0  \n",
       "1                        1                      0  \n",
       "2                        1                      0  \n",
       "3                        1                      0  \n",
       "4                        1                      0  \n",
       "\n",
       "[5 rows x 40 columns]"
      ]
     },
     "execution_count": 571,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Opening the train_flights_subset_cleaned_version_4 to drop some columns and combine with our new 'predicted_flight_delays' colum\n",
    "\n",
    "train_flights_subset_cleaned_version_4 = pd.read_csv('train_flights_subset_cleaned_version_4.csv', sep= \",\")\n",
    "train_flights_subset_cleaned_version_4.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 572,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mkt_unique_carrier</th>\n",
       "      <th>mkt_carrier</th>\n",
       "      <th>mkt_carrier_fl_num</th>\n",
       "      <th>op_unique_carrier</th>\n",
       "      <th>tail_num</th>\n",
       "      <th>op_carrier_fl_num</th>\n",
       "      <th>origin</th>\n",
       "      <th>dest</th>\n",
       "      <th>crs_elapsed_time</th>\n",
       "      <th>flights</th>\n",
       "      <th>...</th>\n",
       "      <th>destination_state</th>\n",
       "      <th>origin_dest_pair</th>\n",
       "      <th>dep_time_of_day_afternoon</th>\n",
       "      <th>dep_time_of_day_evening</th>\n",
       "      <th>dep_time_of_day_morning</th>\n",
       "      <th>dep_time_of_day_night</th>\n",
       "      <th>arr_time_of_day_afternoon</th>\n",
       "      <th>arr_time_of_day_evening</th>\n",
       "      <th>arr_time_of_day_morning</th>\n",
       "      <th>arr_time_of_day_night</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3153</td>\n",
       "      <td>19</td>\n",
       "      <td>2100</td>\n",
       "      <td>3153</td>\n",
       "      <td>SYR</td>\n",
       "      <td>PHL</td>\n",
       "      <td>84</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>PA</td>\n",
       "      <td>SYR-PHL</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>114</td>\n",
       "      <td>4</td>\n",
       "      <td>1997</td>\n",
       "      <td>114</td>\n",
       "      <td>OAK</td>\n",
       "      <td>JFK</td>\n",
       "      <td>319</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>NY</td>\n",
       "      <td>OAK-JFK</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>3276</td>\n",
       "      <td>18</td>\n",
       "      <td>3831</td>\n",
       "      <td>3276</td>\n",
       "      <td>CKB</td>\n",
       "      <td>IAD</td>\n",
       "      <td>64</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>DC</td>\n",
       "      <td>CKB-IAD</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>283</td>\n",
       "      <td>23</td>\n",
       "      <td>3806</td>\n",
       "      <td>283</td>\n",
       "      <td>BWI</td>\n",
       "      <td>CVG</td>\n",
       "      <td>105</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>OH</td>\n",
       "      <td>BWI-CVG</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1021</td>\n",
       "      <td>23</td>\n",
       "      <td>3001</td>\n",
       "      <td>1021</td>\n",
       "      <td>IND</td>\n",
       "      <td>DEN</td>\n",
       "      <td>165</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>CO</td>\n",
       "      <td>IND-DEN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6970</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>660</td>\n",
       "      <td>1</td>\n",
       "      <td>443</td>\n",
       "      <td>660</td>\n",
       "      <td>MIA</td>\n",
       "      <td>PHL</td>\n",
       "      <td>171</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>PA</td>\n",
       "      <td>MIA-PHL</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6971</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1984</td>\n",
       "      <td>1</td>\n",
       "      <td>438</td>\n",
       "      <td>1984</td>\n",
       "      <td>PHL</td>\n",
       "      <td>MIA</td>\n",
       "      <td>186</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>FL</td>\n",
       "      <td>PHL-MIA</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6972</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3337</td>\n",
       "      <td>17</td>\n",
       "      <td>2350</td>\n",
       "      <td>3337</td>\n",
       "      <td>CLT</td>\n",
       "      <td>HPN</td>\n",
       "      <td>115</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>NY</td>\n",
       "      <td>CLT-HPN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6973</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>445</td>\n",
       "      <td>1</td>\n",
       "      <td>414</td>\n",
       "      <td>445</td>\n",
       "      <td>SMF</td>\n",
       "      <td>CLT</td>\n",
       "      <td>273</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>NC</td>\n",
       "      <td>SMF-CLT</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6974</th>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3050</td>\n",
       "      <td>18</td>\n",
       "      <td>2532</td>\n",
       "      <td>3050</td>\n",
       "      <td>MSP</td>\n",
       "      <td>CVG</td>\n",
       "      <td>118</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>OH</td>\n",
       "      <td>MSP-CVG</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>6975 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      mkt_unique_carrier  mkt_carrier  mkt_carrier_fl_num  op_unique_carrier  \\\n",
       "0                      0            0                3153                 19   \n",
       "1                      2            2                 114                  4   \n",
       "2                      8            8                3276                 18   \n",
       "3                     10           10                 283                 23   \n",
       "4                     10           10                1021                 23   \n",
       "...                  ...          ...                 ...                ...   \n",
       "6970                   0            0                 660                  1   \n",
       "6971                   0            0                1984                  1   \n",
       "6972                   0            0                3337                 17   \n",
       "6973                   0            0                 445                  1   \n",
       "6974                   3            3                3050                 18   \n",
       "\n",
       "      tail_num  op_carrier_fl_num origin dest  crs_elapsed_time  flights  ...  \\\n",
       "0         2100               3153    SYR  PHL                84        1  ...   \n",
       "1         1997                114    OAK  JFK               319        1  ...   \n",
       "2         3831               3276    CKB  IAD                64        1  ...   \n",
       "3         3806                283    BWI  CVG               105        1  ...   \n",
       "4         3001               1021    IND  DEN               165        1  ...   \n",
       "...        ...                ...    ...  ...               ...      ...  ...   \n",
       "6970       443                660    MIA  PHL               171        1  ...   \n",
       "6971       438               1984    PHL  MIA               186        1  ...   \n",
       "6972      2350               3337    CLT  HPN               115        1  ...   \n",
       "6973       414                445    SMF  CLT               273        1  ...   \n",
       "6974      2532               3050    MSP  CVG               118        1  ...   \n",
       "\n",
       "      destination_state  origin_dest_pair  dep_time_of_day_afternoon  \\\n",
       "0                    PA           SYR-PHL                          1   \n",
       "1                    NY           OAK-JFK                          0   \n",
       "2                    DC           CKB-IAD                          0   \n",
       "3                    OH           BWI-CVG                          0   \n",
       "4                    CO           IND-DEN                          0   \n",
       "...                 ...               ...                        ...   \n",
       "6970                 PA           MIA-PHL                          0   \n",
       "6971                 FL           PHL-MIA                          0   \n",
       "6972                 NY           CLT-HPN                          0   \n",
       "6973                 NC           SMF-CLT                          0   \n",
       "6974                 OH           MSP-CVG                          0   \n",
       "\n",
       "      dep_time_of_day_evening  dep_time_of_day_morning  dep_time_of_day_night  \\\n",
       "0                           0                        0                      0   \n",
       "1                           1                        0                      0   \n",
       "2                           0                        1                      0   \n",
       "3                           0                        1                      0   \n",
       "4                           0                        1                      0   \n",
       "...                       ...                      ...                    ...   \n",
       "6970                        0                        1                      0   \n",
       "6971                        0                        1                      0   \n",
       "6972                        0                        1                      0   \n",
       "6973                        1                        0                      0   \n",
       "6974                        0                        1                      0   \n",
       "\n",
       "      arr_time_of_day_afternoon arr_time_of_day_evening  \\\n",
       "0                             1                       0   \n",
       "1                             0                       0   \n",
       "2                             0                       0   \n",
       "3                             0                       0   \n",
       "4                             0                       0   \n",
       "...                         ...                     ...   \n",
       "6970                          0                       0   \n",
       "6971                          1                       0   \n",
       "6972                          1                       0   \n",
       "6973                          0                       0   \n",
       "6974                          0                       0   \n",
       "\n",
       "     arr_time_of_day_morning arr_time_of_day_night  \n",
       "0                          0                     0  \n",
       "1                          1                     0  \n",
       "2                          1                     0  \n",
       "3                          1                     0  \n",
       "4                          1                     0  \n",
       "...                      ...                   ...  \n",
       "6970                       1                     0  \n",
       "6971                       0                     0  \n",
       "6972                       0                     0  \n",
       "6973                       1                     0  \n",
       "6974                       1                     0  \n",
       "\n",
       "[6975 rows x 28 columns]"
      ]
     },
     "execution_count": 572,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dropping columns\n",
    "\n",
    "train_flights_sample_submission = train_flights_subset_cleaned_version_4.drop(['origin_airport_id', 'dest_airport_id', 'actual_elapsed_time', 'air_time', 'arr_delay', 'arr_delay_mean_by_mkt_unique_carrier', 'arr_delay_mean_by_op_unique_carrier', 'arr_delay_encoded', 'quarter_Q1', 'quarter_Q2', 'quarter_Q3', 'quarter_Q4'], axis=1)   \n",
    "train_flights_sample_submission                                                                                                                                                           "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One-hot encoding and adding the 'predicted_flights_delay''\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "======================================================================\n",
    "## Stretch Tasks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Multiclass Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The target variables are **CARRIER_DELAY, WEATHER_DELAY, NAS_DELAY, SECURITY_DELAY, LATE_AIRCRAFT_DELAY**. We need to do additional transformations because these variables are not binary but continuos. For each flight that was delayed, we need to have one of these variables as 1 and others 0.\n",
    "\n",
    "It can happen that we have two types of delays with more than 0 minutes. In this case, take the bigger one as 1 and others as 0."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Binary Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The target variable is **CANCELLED**. The main problem here is going to be huge class imbalance. We have only very little cancelled flights with comparison to all flights. It is important to do the right sampling before training and to choose correct evaluation metrics."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base_env",
   "language": "python",
   "name": "base_env"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}