dze-shalom/AirSense_IndabaX_Hackathon_2026

AI-powered air quality monitoring system for Cameroon's 10 regions, predicting pollution, calibrating alerts to local climate realities, and putting actionable insights in the hands of health officers. IndabaX Cameroon 2026 Hackathon. # AirSense Cameroon — IndabaX 2026 > **AI for Climate and Health Resilience in Cameroon** > Real-time PM2.5 monitoring, 7-day forecasting, and health advisory across 85+ cities in all 10 regions. **Live Dashboard →** | **API Docs →** --- ## Table of Contents 1. Overview 2. Architecture 3. Quick Start 4. Progressive Web App (PWA) 5. Data Pipeline 6. Model Training 7. Public API 8. Dashboard Pages 9. Configuration 10. Deployment 11. Innovations 12. Team --- ## Overview AirSense Cameroon predicts PM2.5 air quality across Cameroon using 6 years of satellite reanalysis (CAMS) combined with meteorological data (Open-Meteo). The system: - **Predicts** daily PM2.5 concentrations for 85+ cities across all 10 regions - **Forecasts** 7 days ahead using XGBoost + live Open-Meteo weather data - **Alerts** with Platt-calibrated exceedance probability (F1 = 0.847 at P=0.50) - **Explains** predictions via region-specific SHAP feature attributions - **Projects** future air quality under CMIP6 SSP2-4.5 and SSP5-8.5 climate scenarios - **Advises** schools, farms, and vulnerable groups in English and French - **Reports** downloadable PDF air quality reports per city - **Chats** via a conversational AI assistant powered by Groq (Llama 3.1, free tier) --- ## Architecture ``` AirSense_IndabaX_Hackathon_2026/ │ ├── dashboard/ # Streamlit web application │ ├── app.py # Entry point — routing only │ ├── config.py # Constants: colours, cities, translations, nav │ ├── manifest.json # PWA manifest for mobile install │ ├── components/ │ │ ├── sidebar.py # Navigation, CSS injection, light/dark theme │ │ ├── ui.py # card(), sec(), info_box(), SVG gauges │ │ └── charts.py # PLO() — theme-aware Plotly layout defaults │ ├── pages/ │ │ ├── overview.py # National choropleth map + city rankings │ │ ├── explorer.py # Foreca …