SouthWest Transit App

The agency ran a good service and had no way to tell anyone about it. Schedules were printed, buses ran early or late for ordinary reasons, and every one of those minutes landed on a rider standing outside in February.

Dispatch knew where the buses were. That information lived in a system with no way out of it, so the people who needed it most — the riders, and the dispatchers fielding their calls — were the only two groups who couldn't see it.

Started at the stop. Two weeks riding routes, standing at shelters and sitting in the dispatch office, watching what people actually did while they waited. The pattern was consistent: nobody wanted a map, a schedule or an account. They wanted one number and they wanted to trust it.

So the app opens on that number. Nearest stop, next three buses, minutes. Everything else — routes, alerts, trip planning — sits one layer down, where it doesn't tax the ninety percent of visits that are the same question.

Behind it: a prediction service reading vehicle positions, a dispatcher console that surfaces the same data the riders see, and an alerts tool a supervisor can use with one hand while holding a radio in the other.

SouthWest Transit runs SW Prime on the Spare platform today. What I built has been replaced, which is what happens to transit software over a decade. The work below is history, not a live system.

Adoption came from the shelters, not from marketing. Call volume to dispatch dropped enough that the supervisors noticed before the report did, and the arrival API ended up feeding the printed-schedule replacements at the busiest stops.