BusWatch — RTL Speed Audit

Crowdsourced speed audit of Greater Malé public transport. Riders record trips on their own phones; recorded speed is compared against per-road speed limits; aggregates across many trips build the evidence base. Built from RTL-Speed-Audit-PRD.md (Phase 1 PWA + cloud sync).

Run

npm install
npm run dev        # http://localhost:3008
npm run build      # static export (never run while `next dev` is running — shared .next)

Try it without a bus: tick Demo mode on the home screen — it replays a route with simulated GPS, speeding periodically.

What it does

• RTL Bus — tag one of the 11 RTL Greater Malé routes (Routes 1–5, 8–12, Vilimalé), optional fleet number.
• Taxi / Other (motorbike, joy ride…) — no route tag; the trip logs wherever the vehicle goes.
• Logs GNSS at 1 Hz (fixes worse than 20 m accuracy dropped, < 2.5 m/s treated as stationary).
• Every fix is judged against the limit for the road it's on: nearest corridor segment within 50 m, else the island zone (Malé 30, Hulhumalé 30, Vilimalé 25). Over-limit = speed > limit + 5 km/h margin.
• Offline-first: fixes buffer in IndexedDB; trips upload when online (if sync is configured).
• Privacy: pseudonymous contributor id, first/last 60 s of every trip deleted, per-trip delete, write-only cloud key.

⚠ Limit dataset is placeholder

Geometry and limits in lib/route.ts are approximate. Only the Sinamalé Bridge 50 km/h is publicly reported; everything else awaits the field survey (PRD 7.3). Do not publish findings from this dataset.

Cloud sync (Neon + Vercel)

Trips upload through /api/sync (a Vercel function) into Neon Postgres — the phone never holds a database credential. Tables are auto-created on the first upload.

1. Deployed on Vercel; connect the Neon integration to the project so DATABASE_URL is injected (Vercel dashboard → Storage/Integrations → Neon).
2. For local dev, copy .env.local.example to .env.local with your Neon connection string.
3. Run the views in db/schema.sql once in the Neon SQL editor for trend queries: segment_rates (per-segment speeding rate) and hourly_rates (by hour of day and vehicle), both excluding demo trips.

Without a DATABASE_URL the app still works fully offline; the trips page reports "Cloud sync: off".

Roadmap (PRD)

Phase 2: map-matching validation, analyst dashboard. Phase 3: Capacitor wrapper for background logging. Phase 4: remaining routes, taxi attribution.