Resume Screener

This is a Python program that uses natural language processing techniques to match resumes with job descriptions. The program takes a job description in the form of a Word document and a resume in the form of a separate Word document, and then uses cosine similarity to determine how well the resume matches the job description.

The match percentage is displayed on a gauge chart, and the predicted job position for the resume is displayed. The application also includes a feature to upload multiple resumes for screening and displays the match percentage and predicted job position for each resume on a chart. The application uses the sklearn library for text processing, nltk for tokenization, docx2txt for reading word documents, and streamlit for building the web app. It also includes a class JobPredictor that predicts the job position of the given resume using a trained model saved as pickles.

To run working Notebook in Kaggle

Prerequisites

Before running this program, you will need to install the packages using following command:

pip install -r requirement.txt

How to use

• Clone this repository to your local machine.
• Install the required packages (see Prerequisites above).
• Place the job description Word document in the project directory and name it "temp_jd.docx".
• Run the program using

  streamlit run resume_screener.py.

• Use the file uploader to upload your resume.
• Click the "Submit" button to see how well your resume matches the job description.

Explanation of code

The program uses the docx2txt library to convert the Word documents to text, and the CountVectorizer class from scikit-learn to create a sparse matrix of word counts for each document. The program then uses cosine similarity from scikit-learn to compare the two sparse matrices and compute a similarity score.

The program also uses the pickle library to load pre-trained models for label encoding, word vectorization, and classification.

Finally, the program uses the streamlit library to create a web application with a user interface for uploading the job description and resume documents.

Author

This program was created by Lakpa Sherpa.