Campus bus tracking
Real-time bus arrivals, measured in the open.
Dwell rebuilds campus bus predictions from the raw GPS and publishes its accuracy against an open industry benchmark. Bring your GTFS feed, or build one free in your browser.
- Predictions scored against real arrivals, every weekday
- 63,000
The whole lifecycle
Dwell builds every stage itself, from the schedule and the raw GPS to the rider's phone, and each stage improves the next. Point at a stage to see it; open it for the full story.
And it loops: each trip adds to the timing history and the accuracy record that the next prediction draws from. See how it works →
Bring the GTFS feed you publish, or build one in the editor; it is checked and repaired before anything reads it.
Your schedule: the full stage →As the receiver reports it
Five fixes from one bus, a few seconds apart, where the GPS receiver put them: beside the road as often as on it.
Snapped to the route
The same five fixes on the published line, which is what tells the engine how far along the route the bus is and which stop is next.
Predictions start from the position, speed and heading each bus reports.
Capture raw GPS: the full stage →As the receiver reports it
Five fixes from one bus, a few seconds apart, where the GPS receiver put them: beside the road as often as on it.
Snapped to the route
The same five fixes on the published line, which is what tells the engine how far along the route the bus is and which stop is next.
Each GPS point is map-matched onto the route's real shape, which is what reveals the next stop.
Snap to the route: the full stage →Every hop between stops and every wait at one becomes an observation, grouped by kind of day and part of the day.
Store the history: the full stage →Live position meets learned timing, published as standards-based GTFS-Realtime.
Predict & publish: the full stage →
The live web map under your name and colours, and native apps in the same branding.
Riders: the full stage →Draw a detour, and the closed stops, the new path and the rider alert publish together.
Detours and alerts: the full stage → Operators see their schedule adherence and spacing; dispatch sees every bus on one screen.
Drivers and dispatch: the full stage →
Predictions are compared with when the bus arrived and scored against an open benchmark.
Measure the accuracy: the full stage →Build a GTFS feed in your browser
Seven steps in GTFS's own order, a map of what you are editing, drive times measured along the route, a timetable that fills itself in, and live validation against the same checks Dwell runs before publishing. Import the feed you have or start from scratch. Nothing is uploaded; the network stays in your browser.
What your university gets
Arrival times are rebuilt from each bus's reported position, using timing learned from past trips on your campus.
Riders pick their university and get your colors and logo.
Dwell publishes standard GTFS and GTFS-Realtime feeds, the same formats transit apps read.
Prediction accuracy is scored against the open Transit/Swiftly benchmark and published for riders and administrators to see.
Bring Dwell to your campus
Send a note about your campus, your routes and what tracks your buses today. You will hear back from the person who builds Dwell, usually within a day, with how it would fit your system and what it would cost.