Gulyak.
All work
Web2024

Fleetwise Dashboard

Real-time fleet tracking handling 10k+ daily route updates with sub-100ms latency.

12 weeks4 engineersClient: Logistics Startup
Fleetwise Dashboard

The Challenge

The company was running a fleet of 300+ delivery vehicles across three cities. Dispatchers tracked positions in a spreadsheet updated every 30 minutes — meaning late deliveries were discovered after the fact. Drivers had no in-app way to flag incidents or request rerouting. The existing vendor's SLA was 99.2% — not good enough for same-day delivery SLAs.

Our Approach

We designed and built a full-stack real-time dashboard from scratch. Vehicle telemetry is pushed via WebSockets from GPS devices, ingested into a Redis pub/sub layer, then persisted to PostgreSQL with a 10-second resolution. Dispatchers see a live Mapbox map with colour-coded vehicle states, ETA predictions, and incident flags. Drivers get a minimal React Native companion app for status updates and reroute requests.

  • WebSocket gateway with Redis pub/sub — 10k+ concurrent events/day, p99 < 80ms
  • Mapbox GL JS map with clustering, custom layer rendering, and live polyline updates
  • ETA engine using historical traffic patterns per route segment (PostGIS)
  • Incident reporting flow: driver flags → dispatcher notified < 3s → auto-reroute suggestion
  • Role-based access: dispatcher, ops manager, driver — separate views per role
We went from finding out about late deliveries at 6pm to catching them at 2pm and rerouting. That's the difference between an angry client and a happy one.

Head of Dispatch, Fleetwise

Results

10k+
Route updates per day
<80ms
p99 event latency
99.97%
Dashboard uptime (3 months)
22%
Reduction in late deliveries

Tech stack

ReactWebSocketsRedisPostgreSQL

Have a similar project?

We'd love to hear about it. Free consultation, no strings attached.