AI-powered crime mapping and neighbourhood intelligence
Real estate professionals lacked granular, real-time neighbourhood risk data at the point of purchase or listing decision. Existing tools were either too broad (city-level) or not Canadian.
Real estate professionals lacked granular, real-time neighbourhood risk data at the point of purchase or listing decision. Existing tools were either too broad (city-level) or not Canadian.
Built real-time ingestion from public crime databases, geospatial indexing, and a machine learning risk scoring model that outputs neighbourhood safety scores at the block level.
Live platform serving Canadian real estate professionals and buyers. AI risk scoring operating in real-time across 12 cities.
Data pipeline ingests public crime report feeds across 12 Canadian cities. Geospatial indexing with PostGIS. ML risk model trained on historical incident data weighted by recency, type, and proximity. React frontend with Mapbox for the map interface.
Public crime data quality varies dramatically by municipality. Building a normalization layer that accounts for reporting inconsistencies was the most technically challenging part — and the most important for accuracy.