“Our system is a geospatial landslide early-warning platform that combines static susceptibility factors with dynamic environmental triggers. We collect rainfall, terrain, satellite, soil, geological, hydrological and land-use information, align thern to a common spatial grid, and engineer 14 features including slope, elevation, short- and long-term rainfall, soil, lithology, vegetation, SAR-derived conditions and proximity features. Historical landslide inventories provide supervised labels for a tree-based ML model that estimates location-specific risk. SHAP is then used to explain the major contributors to each prediction. FastAPI exposes the prediction and geospatial services, while a React-based dashboard visualizes the risk on an interactive map and communicates alerts through telegram. The important point is that the system is decision support: it continuously updates risk from the latest available data, but does not claim to guarantee or precisely time a landslide.