TransitStats

A transit trip tracker for enthusiasts to analyze and visualize personal ridership.

Problem

Transit riders have almost no personal tracking tools built for transit. There is still no simple way to build and explore a detailed record of routes, stops, and travel patterns over time.

Features

• SMS + MMS Logging: Start and end trips by text, stop code, stop name, direction shorthand, or stop-sign photo, with route- and direction-aware stop matching.
• History + AI Queries: Build a personal trip history over time, visualize ridership patterns, and ask natural-language questions about your rides.
• Predictions + Journey Intelligence: Get route-aware destination guesses and automatic journey linking that use learned trip patterns without ignoring the physical transit network.
• Transit-Specific Learning: Each completed trip teaches the system stop sequences, travel times, route-stop service, and transfer connections, improving the app over time.
• Public Profiles: Share aggregate ridership stats and streaks through a public-facing profile page.

Stack

• Frontend: Vanilla JS (ES modules), Vite, Leaflet
• Backend: Firebase Auth, Firestore, Hosting, and Cloud Functions
• Messaging: Twilio SMS/MMS
• AI + Models: Google Gemini, V4 logistic regression models, and V5 XGBoost models running through ONNX Runtime
• Transit Intelligence: Custom Prediction, Network, Transfer, and Habit engines

---

• Roadmap
• Changelog
• Security
• Documents

Created by Ryan Hanna | ryanisnota.pro

Powered by Atlas