FloodSense uses historical GIS data, real-time weather APIs, and drain sensor readings to predict hyper-local flooding 6 hours in advance for Delhi's most vulnerable colonies. Push alerts go to local ward councillors and residents over SMS and WhatsApp. The model was trained on 8 years of NDMA data and achieved 87% precision in backtesting.
Directly targets Urban Flooding track of India Innovates Urban Solutions Challenge. Prototype is deployable.
