home page
dashboard
real world location
Analytics
feed back
Citizen Report
report
Features
head trend

Inspiration

Every monsoon season, the same pattern repeats: a city-wide flood alert goes out, and most residents shrug it off because "it's always like this." Meanwhile, someone on a specific street — one that floods every single time because of a poorly graded underpass — gets no warning at all beyond the same generic city-wide notice everyone else received. We kept coming back to a simple observation: disasters are hyperlocal, but alerts are not. A flash flood doesn't care about district boundaries, and neither does a localized fire risk in a dense settlement or a heat-stress pocket with no tree cover. We wanted to build something that treated streets, not cities, as the unit of prediction — and we chose the Disaster Prediction & Alert Systems theme because it let us combine real technical depth with something that could genuinely save someone a flooded commute, or worse.

StreetWatch is an AI-powered early warning system designed to close a critical gap in urban disaster management: the mismatch between city-wide alerts and hyperlocal risk. While existing systems broadcast generic warnings across entire districts, real disasters like flash floods, localized fires, and heat-stress events occur at the scale of a single street or neighborhood. StreetWatch fuses open weather and rainfall data, terrain and elevation maps, historical incident records, and real-time crowd-sourced reports from citizens into a lightweight machine learning model that continuously scores risk across 500-meter grid cells covering the city. When risk crosses a critical threshold in a specific zone, only the residents in that zone receive a targeted SMS or WhatsApp alert, while municipal authorities monitor a live, city-wide risk heatmap to direct emergency response where it's actually needed. By combining predictive modeling with human-in-the-loop reporting, and by focusing on frequent, high-impact micro-disasters rather than rare mega-events, StreetWatch reduces alert fatigue, eliminates sensor blind spots, and remains fully buildable using open datasets — offering a realistic, scalable path from hackathon prototype to real-world civic deployment.