Hi HN! I built Tranzia because I always wondered: “What’s the safest route to take at night?”

Tranzia scores routes (0–10) using official crime data, nighttime visibility, walking exposure, station safety history, and anonymized user feedback. It currently supports NYC and London.

Under the hood it uses H3 spatial indexing, percentile-normalized crime distributions per city, and a transparent scoring formula (55% crime, 20% time of day, 15% walking, 10% feedback). Every route includes a breakdown of why it scored that way.

Would love feedback on the model, data normalization, and extending this to more cities. Live demo here.