All projects
AI Projects · AI-powered geospatial health surveillance platform
Disease Early Warning System
Built an AI-driven disease prediction model combining satellite imagery, climate data, and epidemiological records to forecast disease outbreaks across South Asia.
This flagship project integrates multi-source geospatial data — including MODIS land surface temperature, precipitation indices, and population density layers — with machine learning models to predict cholera and dengue outbreaks 2-4 weeks in advance. Deployed at icddr,b for operational early warning.
Highlights
Traditional disease surveillance relies on lagging indicators. We needed real-time spatial risk assessment that could predict outbreaks before they happen.
Combined satellite-derived environmental variables with historical case data using ensemble ML models (Random Forest + LSTM), served through a PostGIS-backed web dashboard.