Resume Analyzer is a prototype web application that allows users to upload multiple resumes and compare them against a job description using vectorization and cosine similarity. The project is built using Python, Flask, and scikit-learn. It can also be enhanced using machine learning models for better efficiency and accuracy — work on this is in progress.
- Supports multiple resume formats:
.pdf,.docx,.txt - Uses TF-IDF vectorization and cosine similarity for scoring
- Ranks resumes based on relevance to the provided job description
- Allows users to upload and store resumes in a custom local path
- HTML/CSS templates can be customized as per your needs
- Can be deployed on the web using services like PythonAnywhere
- Users input a job description and upload up to 10 resumes.
- The application extracts text from all resumes:
- PDF: extracted using PyPDF2
- DOCX: extracted using docx2txt
- TXT: extracted via standard file read
- Text data is vectorized using
TfidfVectorizerfrom scikit-learn. - Cosine similarity is calculated between the job description and each resume.
- The top 3 matching resumes are displayed along with their similarity scores.
- Python 3.7 or higher
- Required Python packages (Check Requirements.txt)
git clone https://github.com/your-username/resume-analyzer.git
cd resume-analyzer
pip install -r requirements.txtpython main.pyResume files will be stored in the uploads/ directory. You can change this path in the main.py file as needed.
This app can be deployed to web platforms such as:
- PythonAnywhere
- Render
- Railway
- Heroku (with appropriate setup like a
Procfile)
Work is ongoing to enhance the application using machine learning for smarter resume-job matching. Planned features include:
- Deep learning-based similarity scoring
- Better handling of resume formatting and structure
- Filter options for specific qualifications, skills, or keywords
- User accounts and dashboard
Contributions are welcome.
- Frontend: HTML, CSS (with customization options)
- Backend: Python Flask
- Text Processing: TF-IDF, Cosine Similarity
- File Parsing: PyPDF2, docx2txt
This is a prototype and may not handle every edge case. It serves as a proof-of-concept for resume screening and ranking based on textual similarity. For production use, improvements in model robustness and user interface are recommended.
MIT License