Job Recommendation System Using LinkedIn Data

A machine learning–powered system that recommends jobs based on user-input skills using a cleaned LinkedIn dataset.

Dataset

• Source: Kaggle – LinkedIn Jobs & Skills 2024
• Used:
  • linkedin_job_postings.csv
  • job_skills.csv

Project Overview

The system tackles the problem of information overload in job search by recommending roles based on skill matching.

Key Phases

• Phase 1: Data cleaning & EDA (handled missing data, merged datasets, visual insights)
• Phase 2: Clustering (KMeans) & model training (SVM, Logistic Regression, etc.)
• Phase 3: Deployed via Flask Web App

Models Used

• Logistic Regression
• Decision Tree
• K-Nearest Neighbors
• Naive Bayes
• Random Forest
• XGBoost
• ** Best Accuracy: Support Vector Machine (76%)**

Web App (Flask)

Features

• Input: User enters skills (e.g., nursing, coding)
• Output: Top 5 job matches with title, location, and LinkedIn URL
• Built with Flask + HTML/CSS

How to Run

# Setup
python -m venv venv
source venv/bin/activate  # or .\venv\Scripts\activate on Windows
pip install -r requirements.txt

# Run the app
python app.py