DevLucks/Predicting-Climate-driven-diseases

Real-time disease outbreak surveillance for Nigeria — predicts Cholera & Lassa Fever risk using ML (XGBoost) and live climate data # Outbreak Watch — Climate-Driven Disease Surveillance System A real-time disease outbreak surveillance platform for Nigeria that predicts **Cholera** and **Lassa Fever** risk using climate data, ML ensemble models, and live weather feeds. **Live Demo:** outbreak-watch-frontend.onr… --- ## Overview Outbreak Watch combines machine learning with real-time weather data to predict disease outbreak risk across Nigerian states. The system achieves **74% prediction accuracy** using an ensemble of XGBoost and scikit-learn models trained on historical climate and disease incidence data. ## Features - Real-time risk scoring for Cholera and Lassa Fever across Nigerian states - Interactive map and 3D globe visualization of outbreak risk levels - Live weather data integration (temperature, humidity, wind) via Open-Meteo API - Historical disease trend charts powered by WHO GHO data - ML ensemble model with 74% accuracy on held-out test data - Fully responsive — works on desktop and mobile --- ## Tech Stack | Layer | Technology | |-------|------------| | Frontend | React 18, TypeScript, Vite | | Visualization | Recharts, Leaflet, react-globe.gl, Three.js | | Animation | Framer Motion | | Backend | FastAPI (Python) | | ML Models | scikit-learn, XGBoost | | Data Processing | pandas, numpy, joblib | | Weather API | Open-Meteo (no API key required) | | Disease Data | WHO GHO API | --- ## Project Structure ``` Final-Year-Disease-Prediction/ ├── backend/ # FastAPI + ML inference service │ ├── main.py # App entry point, CORS, routes │ ├── model_service.py # Loads .pkl models, exposes predict() │ ├── weather_service.py # Fetches live weather from Open-Meteo │ ├── disease_service.py # Fetches cholera data from WHO GHO API │ └── requirements.txt ├── frontend/ # React + TypeScript + Vite │ └── src/ # Components, pages, charts, map ├── notebooks/ # Data explorati …