RailCrowd

RailCrowd is a mobile + web app that predicts how crowded each light-rail train and even each car will be, 5–30 minutes ahead, so riders can choose the least packed option or adjust departure time. It combines official GTFS-Realtime feeds (where available) with anonymous, opt-in rider signals (quick “car fullness” taps, passive Bluetooth/Wi‑Fi density estimates where legal, and station-level footfall partners) to build crowding models per line, time, and event patterns. The app shows a simple heatmap by car, alerts when a usually-quiet trip is about to spike, and provides accessibility-aware guidance (e.g., which car aligns with elevators/ramps). This is an AI + traditional app: AI for forecasting and anomaly detection, traditional for routing, alerts, and feed ingestion. Monetization is B2B (transit agencies, venues, employers) plus optional rider subscription for advanced alerts.