Smart Insurance Risk Predictor

AI-powered insurance premium prediction and risk classification using Machine Learning.

🔗 Live Demo: smart-insurance-risk-predictor.vercel.app

Project Structure

├── main.py              # Backend: ML pipeline + Flask API server
├── insurance.csv         # Training dataset
├── models/               # Saved ML models (auto-generated)
├── src/
│   ├── App.tsx           # React frontend UI
│   ├── main.tsx          # React entry point
│   └── index.css         # Styles
├── index.html            # Vite HTML entry
├── requirements.txt      # Python dependencies
├── package.json          # Node.js dependencies
├── vite.config.ts        # Vite configuration
└── tsconfig.json         # TypeScript configuration

Setup & Run

Prerequisites

• Python 3.10+
• Node.js 18+

Backend (Python/Flask)

# Create virtual environment
python -m venv .venv

# Activate (Windows)
.venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Run (trains models + starts API on port 5000)
python main.py

Frontend (React/Vite)

# Install dependencies
npm install

# Start dev server (port 5173)
npm run dev

# Build for production
npm run build

Usage

1. Start the backend: python main.py (runs on http://localhost:5000)
2. Start the frontend: npm run dev (runs on http://localhost:5173)
3. Open http://localhost:5173 in your browser
4. Fill in the form and click Predict Insurance

API

POST /predict

Request:

{
  "age": 35,
  "sex": "male",
  "bmi": 25.0,
  "children": 2,
  "smoker": "no",
  "region": "northwest"
}

Response:

{
  "Predicted Premium": 19108.61,
  "Risk Class": "Low Risk"
}

ML Models

• Regression: Linear Regression, KNN Regressor → best saved automatically
• Classification: Logistic Regression, KNN Classifier, Naive Bayes → best saved automatically

License

This project has no copyright claim but please don't copy it 👉👈.