FuelBuddy Vault is a GPS and telematics dashboard for tracking the delivery fleet: live vehicle positions, route playback, and driving-behavior reports. I was a contributor on the frontend team building it.
What I worked on
- Live fleet map: Worked on the live vehicle map with status and vehicle filters, clustering, and a route-playback view with configurable speed.
- Reporting: Contributed to a custom report builder and an overspeed report, both built on real vehicle-tracker telemetry rather than simulated data.
- Data layer: Worked within a Relay/GraphQL data layer with Recoil for client state.
Why it mattered
Turning raw GPS-tracker data into a live map and usable reports gave FuelBuddy's operations team visibility into fleet location and driving behavior without needing to read hardware telemetry directly.
Real Hardware Data, Not a Mock
The overspeed report's field mapping matches Teltonika's AVL parameter ID conventions exactly (ignition state, movement sensor, total odometer, GNSS speed, and BLE-based temperature and fuel-level sensors), meaning this dashboard consumes genuine telemetry from GPS/OBD tracker hardware installed in the vehicles, not a simulated data feed built for demo purposes.
Data Model & Scale
- Relay's normalized cache and GraphQL fragments back the map and report views, a data-fetching pattern built for a large, changing dataset rather than a handful of fixed queries.
- The report builder supports flexible query composition (date range, asset type, customer asset) instead of a fixed set of canned reports.
Stack
React, Relay, Recoil for state, TanStack Table, Chart.js, Google Maps, Firebase.