insurance-premium-prediction-fastapi

This project is an Insurance Premium Prediction API built with FastAPI. The main objective is not only to provide a working insurance premium prediction service, but also to serve as a blueprint for building industry-grade FastAPI applications. The codebase demonstrates best practices in API design, modular code organization, input/output validation, and production readiness.

Features

• FastAPI: High-performance, easy-to-use Python web framework.
• Modular Structure: Clear separation of models, schemas, and API logic.
• Input Validation: Uses Pydantic for robust request/response validation.
• Prediction Endpoint: /predict endpoint for premium category prediction.
• Health Check: /health endpoint for service monitoring.
• Interactive Docs: Swagger UI available at /docs.
• Error Handling: Graceful error responses for invalid requests or server issues.
• Production Ready: Example .gitignore, clear project structure, and extensible code.

Demo

Below is a demonstration of the working product:

Docker Hub

You can pull the pre-built Docker image from Docker Hub:

https://hub.docker.com/r/anujjainbatu/insurance-premium-api

docker pull anujjainbatu/insurance-premium-api

API Link

Once running (locally or via Docker), access the API at:

• Swagger UI: http://localhost:8000/docs
• Health check: http://localhost:8000/health

Public API Endpoint

If you want to use this API in your own application, you can access the deployed instance at:

• Base URL: http://51.20.55.205:8000/
• Swagger UI: http://51.20.55.205:8000/docs
• Health check: http://51.20.55.205:8000/health

Feel free to integrate this endpoint directly into your projects for insurance premium

Project Structure

insurance-premium-prediction/
│
├── app.py                      # Main FastAPI application
├── model/
│   └── predict.py              # Model loading and prediction logic
│   └── model.pkl               # Serialized ML model (not included in repo)
├── schema/
│   ├── user_input.py           # Pydantic model for input validation
│   └── prediction_response.py  # Pydantic model for output schema
├── venv/                       # Virtual environment (excluded from git)
├── README.md                   # Project documentation
├── .gitignore                  # Git ignore file
└── ...

Getting Started

1. Clone the repository

   git clone <repo-url>
   cd insurance-premium-prediction

2. Create and activate a virtual environment

   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies

   pip install -r requirements.txt

4. Run the API

   uvicorn app:app --reload

5. Access the API

   • Swagger UI: http://127.0.0.1:8000/docs
   • Health check: http://127.0.0.1:8000/health

Example Request

POST /predict

{
  "bmi": 27.5,
  "lifestyle_risk": "medium",
  "age_group": "30-40",
  "city_tier": "tier_1",
  "income_lpa": 12,
  "occupation": "salaried"
}

Response

{
  "premium_category": "Medium",
  "confidence": 0.85,
  "class_probabilities": {
    "Low": 0.1,
    "Medium": 0.2,
    "High": 0.7
  }
}

Notes

• This project is intended as a reference for building scalable, maintainable FastAPI applications in a production setting.

License

MIT License

---

Blueprint for Industry-Grade FastAPI Apps
Use this project as a starting point for your own robust, production-ready APIs!