Blog Feature

GTFS: The promise of data-driven public transit

Standardized transit data is a harbinger of a future where the fastest way to get around doesn’t require a private car.

By Eric Jaffe

Portland's TriMet system

This piece is part of the Sidewalk Talk series “15 Innovations That Shaped the City.”

Public transportation agencies have always struggled to communicate information to riders—an effort historically resigned to a hodgepodge of pamphlets, placards, and websites. All that’s changed over the last decade with the mass adoption of smartphones, alongside mobile apps that can display real-time bus and train locations and forecast accurate arrival times. The secret sauce behind these tools is a humble data format called the General Transit Feed Specification. GTFS not only gives city residents an unprecedented level of visibility into their personal mobility options, it represents the power of open data, digital platforms, and public-private collaboration to enable innovations that improve urban life.

The origins

In 2005, Bibiana McHugh, an IT manager for the Portland, Oregon, transit authority TriMet, went on an investigative road trip. She tried to use public transit in the cities she visited but found the process confusing and frustrating. Navigation apps such as MapQuest and Yahoo Maps were good at cranking out driving directions, but for public transit—even in the age of GPS—McHugh was forced to rely on old-school timetables and street signs.

The root problem: Public transit in the U.S. (and in most large countries) is split among hundreds of local transit authorities, each operating in its own closed network. Despite efforts to create a sprawling federal data standard called TCIP (Transit Communications Interface Profiles), few agencies had adopted it by 2005. When she got back from her trip, a frustrated McHugh began calling technology companies to see if any were interested in helping standardize transit scheduling data. Eventually, she reached Chris Harrelson, an engineer at Google.

The breakthrough

As it happened, Harrelson and a small team at Google had been working on the same problem and hitting the same roadblocks. Most transit agencies were nervous or unwilling to share data or didn’t have the budget. Within a few days, however, McHugh was able to get TriMet’s permission to share the agency’s data with Google. Within months, the experiment produced a powerful result: a new, open-source scheduling-data format called Google Transit Feed Specification, or GTFS. (The G was later changed to stand for “general.”)

Despite the long name, GTFS is very simple: It is a predictably structured list of transit stop locations and times. In software, simplicity equates to usability. By the end of the year, Portland had real-time transit schedules up and running. Five more cities joined in 2006, and within a few years, GTFS was in use by hundreds of transit agencies and outside developers.

The impact

Today, people who live, work, or visit cities can turn to GTFS-powered mobile apps to find transit routes and plan trips. They can also track live locations of buses and trains thanks to two newer data formats: GTFS Realtime, and a competing standard in Europe called Siri (no relation to Apple’s voicebot). Some cities with public bike-share services now use GBFS (the B is for “bike-share”). The most complete apps, such as Transit, pull in both these public sources and data from private mobility services such as Lyft or Uber, showing users a complete array of transit options for each trip.

Though it’s early, these tech-driven efficiencies have already had measurable impacts on city living. Recent studies of the new tools show that riders are enjoying shorter wait times and feeling more satisfied with transit services overall, potentially leading to more ridership and less car reliance. Those benefits stand to increase as real-time tracking becomes more accurate. While GPS can pinpoint live locations, it’s much harder to predict delays, which is why apps such as Transit have begun crowdsourcing data on transit delays in some cities, much like the popular driving app Waze crowdsources traffic.

Today, more than 1,000 transit agencies worldwide rely on GTFS. “GTFS has allowed agencies to become more agile,” says Sean Barbeau, a principal mobile software architect at the University of South Florida’s Center for Urban Transportation Research. “At the beginning, it was just a way to push schedules into the world.” Today, widespread GTFS adoption allows simultaneous cost-cutting and service improvements: now that a simple and ubiquitous standard exists, agencies can forego costly, slow internal processes and work more closely with bordering cities that use the same data format, ensuring continuous service.

The future

McHugh, who is still a manager at TriMet, dreams of a future beyond just buses and trains—one in which transit technology can stitch together almost any desired route. A long trip could advance from a sidewalk to a bike path to a transit center. A bus ride would end at another transit center, where a ride-hail vehicle would be waiting to take you the rest of the way. TriMet and other agencies are working to expand OpenTripPlanner, a service created by transit agencies that collects and standardizes data from many sources, for such multimodal trips.

Many transit agencies, perpetually underfunded and subject to changing demand, are taking a wait-and-see attitude. “There’s a lot of talk about how transit agencies are becoming mobility management centers—how it’s not just about public transportation anymore,” says McHugh. “To oversee it from a higher level and broader perspective is something that is going to be needed and necessary.”

The mix of potential transit choices is exciting for city residents. Even more exciting are the endless mobility possibilities that will arise from the combination of data-driven tools and self-driving vehicles—a partnership that may well give rise to a new era of urban mobility anchored around point-to-point transit. GTFS may best be understood as a harbinger of a transportation future where the fastest way to get around doesn’t require owning a private car.