Predict, visualize, and plan for cleaner air in smart cities using AI for Earth principles.
AIRGUARD is a full-stack application designed to tackle urban air pollution. By integrating smart city sensor data (PM2.5, PM10, NO₂, CO, O₃) with advanced machine learning, the system provides accurate Air Quality Index (AQI) predictions.
This project empowers policymakers and citizens with early warning mechanisms and interactive data visualization to foster healthier urban environments.
The core of AIRGUARD lies in its ensemble of regression models. We evaluated and implemented multiple supervised learning algorithms to ensure the highest predictive accuracy:
- XGBoost (Extreme Gradient Boosting): For high-performance gradient boosted decision trees.
- Gradient Boosting Regressor: To optimize prediction through additive modeling.
- AdaBoost Regressor: For boosting the performance of weak learners.
- Random Forest: An ensemble approach to reduce overfitting and improve stability.
- Decision Tree: To establish a clear, interpretable baseline for data splits.
| Layer | Technology |
|---|---|
| Frontend | React.js, Vite, Tailwind CSS, Recharts |
| Backend | FastAPI (Python), Uvicorn |
| ML Libraries | XGBoost, Scikit-Learn, Pandas, NumPy |
| Environment | Microsoft AI for Earth Principles |
AI-AQI/
├── backend/ # FastAPI & ML Models
│ ├── app.py # API Endpoints
│ ├── models/ # Trained .pkl or .joblib files
│ └── requirements.txt
├── frontend/ # React + Vite
│ ├── src/ # UI Components & Dashboard
│ └── package.json
└── README.md
git clone [https://github.com/your-username/AI-AQI.git](https://github.com/your-username/AI-AQI.git)
cd AI-AQI
cd backend
pip install -r requirements.txt
uvicorn app:app --reload
💡 Tip: View the interactive API documentation at
http://127.0.0.1:8000/docs
cd ../frontend
npm install
npm run dev
💻 Dashboard available at:
http://localhost:8080
- 🔹 Multi-Model Inference: Real-time AQI prediction using XGBoost and Ensemble methods.
- 🔹 Dynamic Dashboard: Visual representation of pollutants (PM2.5, NO₂, etc.) using interactive charts.
- 🔹 Early Warning System: Automated alerts for hazardous air quality levels.
- 🔹 Scalable API: Modular FastAPI backend ready for integration with IoT sensors.
A huge thanks to the team behind AirGuard.
| Contributor | GitHub Profile |
|---|---|
| Shreshth Verma | 👤 @Shreshth1805 |
| Diksha Mahajan | 👤 @diksha-mahajan |
| Pridhi Thareja | 👤 @Pridhi-24 |
| Nipun Mahajan | 👤 @nipunmah |
| Ishwinder Kaur | 👤 @IshwinderKaur8 |
This project is licensed under the MIT License - see the LICENSE file for details.