Ghost buses and the trust gap.

The screen says the bus is three minutes away. Three minutes pass. The bus is still three minutes away, until the prediction vanishes and the next one is twenty minutes out. Or the vehicle never appears at all. Riders call these ghost buses: trips shown in an app or on a display that do not arrive as predicted.

I am Ines Halvorsen, a former transit planner. Real-time information depends on data that can go stale or be incomplete. A GPS device may stop reporting. A vehicle may be pulled from service but remain in the feed. Some predictions may be based on a schedule rather than a live location, without making that distinction clear to the person waiting.

Live is not always live.

A prediction is only as current as the information behind it. Agencies and vendors need to detect when a vehicle has stopped sending location updates, distinguish scheduled service from a vehicle actually being tracked, and reconcile dispatch changes with passenger information. A bus that has been canceled should not remain on the map as if it is approaching.

One ghost bus can do more damage than ten late ones: it asks someone to trust a trip that does not exist.

Late service is frustrating, but a rider can at least see what happened and decide what to do next. A ghost prediction takes away that choice. Someone may wait in bad weather, miss a connection, or turn down another option because the screen says help is coming. The loss is not only time; it is confidence in every future prediction.

Agencies should label information clearly as “live” or “scheduled,” show when a vehicle last reported its location, and avoid presenting an old estimate as current fact. When a trip is canceled or a bus is removed from service, the change should be passed through to the real-time feed promptly. If data is unavailable, say so directly rather than displaying false precision.

The fix is partly technical and partly operational. Someone must own the accuracy of the passenger feed, compare it with dispatch information, and give riders a way to report recurring errors. Agencies should publish service changes consistently across signs, apps, and websites. A clear “no live data” message is less impressive than a countdown, but it is more respectful.

Real-time information is a promise about what the agency knows. Make the boundary between schedule and observation visible, remove trips that will not run, and treat data quality as part of service reliability. Riders can adapt to an honest delay. They should not have to guess whether the bus exists.

The distinction is especially important when an agency publishes a standard transit feed used by many apps. A correction made only in the agency’s own app may not reach the other places riders check. Service alerts and vehicle-location data need to stay consistent across the feed, station displays, and customer channels. When a trip is canceled, make the cancellation travel with the trip information.

Agencies should also make it easy to report a mismatch: route, stop, time, and what the screen showed. A single report may be a temporary device fault; repeated reports can reveal a dispatch or data process that needs attention. The point is not to promise perfect prediction. It is to make the information honest enough that riders can make their own decisions.

— Ines