Raincast

Raincast is a web-based platform that predicts rainfall patterns based on regional, and seasonal data using machine learning models. Built with Next.js (frontend) and FastAPI (backend), it uses a historical rainfall dataset to deliver accurate rainfall predictions for different subdivisions of India.

Features

• Predict rainfall based on month, year, and region
• One-hot encoded location-based prediction (36 subdivisions)
• Trained ML models with accuracy and evaluation metrics
• Backend powered by FastAPI with ML model integration
• API endpoint for real-time prediction

Machine Learning Model

• Model Used: Random Forest Classifier / XGBoost
• Data: Monthly rainfall data (1901–2015) from 36 Indian subdivisions
• Target: RainToday (0 or 1)
• Features:
  • Year & Month
  • Subdivision (One-hot encoded)
  • Seasonal indicators (Monsoon, Summer, Winter, etc.)
  • Monthly rainfall data (JAN–DEC)

Result

Setup Instructions

1. Clone the repository

git clone 
cd raincast

2. Backend – FastAPI

cd rainfall-api
pip install -r requirements.txt
python -m uvicorn app.main:app --reload

3.Frontend - Next JS

cd raincast
npm install
npm run dev

.env

Backend
FASTAPI_API_URL=" "
Frontend
CORS_ORIGINS=[" "]

Run Frontend and backend at the same time