Spatiotemporal Machine Learning-Based Air Quality Forecasting and Health Advisory Platf...

Objective: To engineer a deployed ML system that predicts local air quality levels and issues real-time health advisories using spatiotemporal data for Indian cities.

Air pollution is a persistent challenge in India's urban centers, affecting millions with respiratory and cardiovascular risks, especially during peak seasons and events like crop burning or festivals. Citizens, hospitals, and municipal authorities lack timely, localized forecasts and actionable health advisories to mitigate exposure.

The proposed solution leverages spatiotemporal machine learning to predict air quality indices (AQI) at ward-level granularity, integrating government sensor feeds (CPCB/IQAir), satellite imagery, meteorological data, and mobility patterns. It delivers personalized health advisories and forecast dashboards via a web/mobile app, supporting decision-making for both individuals and public health agencies.

Key features include a working ML pipeline trained on open Indian air-quality datasets, a real-time visualization dashboard, automated health-risk notifications, and deployment on cloud infrastructure. The demonstration will showcase prediction accuracy, live forecast updates, and advisory generation for Mumbai and Delhi regions.

Industry and societal impact includes empowering citizens to reduce exposure, aiding hospitals in resource allocation, supporting policy response, and scalable extension to other cities. The project aligns with India's Smart Cities and public-health digital initiatives.