Your Transit Feed Exists. That Doesn’t Mean Anyone Can Find It.

Connectively

Connectively connects subject-matter experts with top publishers to increase their exposure and create Q & A content.

• 3 min read

Your Transit Feed Exists. That Doesn't Mean Anyone Can Find It.

© Image Provided by Connectively

Written by Nick Sawinyh

Earlier this year I ran a scan across the six Southern California counties served by the regional 511 traveler information system. I took every transit provider there that reports to the Federal Transit Administration’s National Transit Database, then joined that list against the Mobility Database, the open catalog that trip planners and aggregators use to find transit feeds. The question was narrow: if you were a rider, an app developer, or a regional program trying to cover the area, how many of these agencies could you actually get live data from?

108 providers report to NTD across Los Angeles, Orange, Riverside, San Bernardino, Ventura, and Imperial counties. Ten had a real-time feed that was cleanly linked and discoverable in the public catalog. Twenty-three published schedule data in GTFS, the open format nearly every trip planner reads, but with no catalog-linked real-time. The remaining eighty-one showed a gap of some kind: nothing discoverable at all, a feed declared to NTD that could not be matched, a feed gone stale, or a feed that failed validation.

Then came the part that changed how I read the whole exercise. Several of the largest agencies sitting on that gap list run real-time service, and have for years. LA Metro, OCTA, and Long Beach Transit all operate GTFS-Realtime, the live companion to that format. What they have is a registration gap, a staleness gap, or a broken link between the catalog record and the data itself. The information is being produced and published. It simply is not findable through the channels everyone downstream actually queries.

That distinction matters more than the headline count, and it travels well outside transit.

Publishing is an event. Discoverability is a state.

Since report year 2023, fixed-route NTD reporters have been required to maintain a public GTFS feed and give FTA a working link to it. The database is worth understanding on its own terms if you work anywhere near this data, and this primer on what the NTD actually holds is a good place to start. Enforcement runs through reporting and grant eligibility rather than any fine schedule, and FTA’s stated method is to periodically check that the submitted web links are viable and current.

That is precisely the check most agencies never run on themselves. A feed gets stood up during a project, someone confirms it loads, and the box is marked done. Nobody owns the question of whether it still resolves, still validates, and still appears where an outside consumer would look for it. Two years later the URL has moved, the schedule inside has expired, or the catalog entry points somewhere dead, and the agency has no idea, because internally everything looks fine.

Three things I took away

The first is to measure from the outside in. Whatever you publish, score it the way a stranger encounters it, through the public directory or registry your consumers actually use, not through your own server logs. Our gap list changed substantially the moment we stopped asking “does this feed exist” and started asking “can someone find it without calling us.”

The second is that validation decays. A feed that loads is not a feed that passes. Schema errors, expired service dates, and a hosting URL that quietly changed all produce the same outcome for a downstream consumer, and none of them announce themselves. Whatever you checked at launch has to be rechecked on a schedule someone actually owns.

The third is the one I would push hardest. Size the problem before you assume its shape. Going in, I expected the story to be small rural operators with no technical capacity. The actual finding was that the region’s biggest and best-resourced agencies were just as likely to be invisible, for entirely different reasons. Had we built a program around the assumption we started with, we would have aimed it at the wrong half of the list.

None of this is unique to transportation. Any organization that publishes data for others to consume, whether an API, a catalog listing, or a public dataset, has the same exposure: an internal record of having shipped something, and no standing check on whether the outside world can still reach it. The gap between those two states is usually invisible right up until it is expensive.

Author Bio:
Nick Sawinyh is Head of Product and GTM at
Veodyn, which builds federated transportation data infrastructure for public agencies. He works on GTFS and transit data standards, and on how regional data hubs get bought and deployed.

Up Next