When Data Derails Transit: The Invisible Infrastructure Crisis

Your transport system is invisible to Google

Picture a busy airport arrivals area, passengers in long queues for taxis right underneath a bright red "Train to City" banner. This may not seem frustrating and illogical to a transport planner but it’s the result of a breakdown in digital wayfinding and journey planning information.

At base, Google Maps didn’t have the railway station anchored at the same location as the airport terminal, so it de-prioritised the rail option when people searched for journeys to the city. Not to name and shame the airport in question but "Acme Airport" has multiple locations in Google Maps, not tied together as a central transport interchange. As a result, when the passenger searched for ‘Acme Airport’ the algorithm linked this to the road node and thus prioritised the road forecourt as the origin and so the trip planner offered a taxi and ride-share route, not the train station, with its own node, only metres away.

The resulting passenger behaviour may seem strange, especially when Inside the terminal, all the dominant signage pointed to the rail link. But as soon as the passenger left the airport’s welcome area and moved their reliance to their phone, the railway station didn't exist.

The data said one thing. The built environment said another. That mismatch made something clear: if the data describing a system is wrong, the investment beneath it becomes invisible.

The pattern repeats everywhere

This isn't unique to Acme Airport.

Luton built a £225 million railway to its airport. The DART opened in 2023, cutting journey times from St Pancras to 32 minutes. Yet Google Maps still routes many passengers to buses because the station is coded as "Central Terminal Station" rather than being properly associated with the airport.

The engineering is flawless. The metadata is broken.

The pattern is formulaic now. A new rail link opens. Engineering and operations teams deliver to specification. The PR team launches it. Then the data layer gets orphaned.

No single source of digital truth. No one owns the reconciliation between operator timetables, authority boundaries, and tech firm basemaps. The infrastructure exists, but in data terms, it's fragmented.

It's no one's explicit job

Airport stations sit in a grey zone. The rail operator holds the timetable. The airport controls the land. Digital platforms draw their base map from third-party datasets that interpret both.

To fix it, someone has to publish authoritative geospatial data linking the terminal to the station with proper geometry, naming conventions, and service codes.

But there's no mandated feedback loop. Each actor assumes another will correct the metadata. The error persists and replicates across every API that consumes the same base layer.

The sector still treats information as a by-product, not an asset. Project sponsors stop at physical completion. Once the line opens, responsibility for its digital representation dissolves into a patchwork of marketing teams, local authorities, and mapping platforms.

This circle of ownership is further complicated as many airports rely on car parking as an income stream so their will to make the city authority’s train a success may be limited.

AI agents will make this worse

The next generation of travel assistants are probabilistic systems. They blend live data with behavioral inference.

If source feeds are inconsistent or incomplete, the agent doesn't show an error. It improvises. It fills gaps with best guesses, often shaped by commercial partnerships or past user behavior.

That's how you get competing realities. One agent nudges you toward a car-share because it lacks reliable rail data. Another defaults to a flight because the regional coach service never published its feed.

Each system learns from its own skewed reflection of the network.

The transport conversation is still focused on open data compliance while tech platforms are already training agents to generate personalised advice from whatever datasets they can find. The two timelines don't line up.

The fix costs almost nothing

Add one line to every delivery plan: appoint a data steward with authority and budget from day one.

Not an IT manager. Not a communications officer. A named role, written into governance structure, responsible for accuracy, publication, and upkeep of every dataset that defines the service.

They sit alongside safety and operations leads at project board level with formal sign-off rights. When track layout changes or a station name is agreed, data updates trigger automatically.

The cultural impact is huge. It signals that data isn't a by-product. It's part of the asset.

Transport for London does this. They built a single system of record for identifiers, fares, and data relationships. They treat that system as operational infrastructure. When TfL updates a station name, it propagates automatically across journey planners, signs, and retail systems.

The result? Journey planners agree. Tickets behave predictably. Disruption messages line up with reality.

Success looks ordinary. That's the point.

Let’s compare that to some National Rail platforms where stations have different names depending on which source you use, looking at you Edinburgh, Edinburgh (Waverley), Waverley, Edinburgh Waverly, Edinburgh Waverly.

You're already paying for It

Every new rail line already depends on digital representations: stop codes, timetables, fare tables, journey planners. Those datasets exist, but they're treated as paperwork rather than assets.

You're paying for them repeatedly through consultants, contractors, marketing teams, and tech integrators because no one owns them end to end.

London's open transport data generates between £90 million and £130 million in annual value. That's the business case.

The fix costs almost nothing compared with physical spend. Mandate a single data steward per project. Fund its maintenance for the asset's life. Make accuracy auditable like safety, accessibility, or carbon.

You can't deliver 21st-century transport with 20th-century data habits.

Build the digital layer properly, and everything else follows: ridership, accessibility, public trust.