AI-powered landslide risk intelligence for the Northeast.
PRAVAAH combines rainfall, terrain, historical landslide evidence, geospatial intelligence and field observations to help authorities understand risk earlier and act faster.
In the mountainous terrain of the Northeast, rainfall, terrain steepness, historical scars, road blockades, and field reports exist in separate administrative silos.
Result: Authorities react after roads collapse; citizens receive generalized regional alerts too late.
Connects dynamic rainfall with real terrain physics to predict sector failure probabilities, reveal SHAP drivers, simulate surges, and compute safe detours.
Four scientifically grounded capabilities bridging machine learning and civil defense.
Estimate landslide risk from environmental and terrain conditions using calibrated XGBoost inference.
See which factors are driving the model's risk estimate via TreeSHAP feature contributions and local physics.
Test how changing rainfall conditions (+20%, +40%, +100%) alter hazard scores and civil mandates.
Prioritize roads, settlements and critical infrastructure via multi-criteria ranking and safe detour routing.
Citizens receive simple, actionable guidance. Authorities access full geospatial command tools. Field scouts capture verifiable evidence.
For District Collectors, DDMA, and BRO Highway Engineers
For Geological Field Scouts, Ward Wardens & Volunteers
For Local Residents, Highway Commuters & Tourists