Open data standards have transformed real-time passenger information and transit technology.

Transit agencies today can invest in data quality and accuracy and, by ensuring their hardware and software support open data, have confidence in data consistency and accessibility across their operations.

The GTFS standard is the foundation of modern transit data. The standard, developed jointly by TriMet and Google, spearheaded the movement towards open data standards and opened the door to the modern transit rider experience.

Bibiana McHugh serves as the Mobility and Location-Based Services IT Manager at TriMet. She played a foundational role in producing the GTFS standard and in subsequent data standard evolutions such as GTFS-rt.

Our General Manager, Transit, Ritesh Warade, sat down with McHugh after APTA Mobility in Portland for a conversation on the origins of GTFS and where data standards are headed in the future.

‍

Dive in for transit technology insights and never-before-told anecdotes from the early days of transit data standards.

‍

RW: How did you get started or involved with transit data? What was the spark?

BM: I have this Kansas City Union Station timetable from 1917 and it's one of my favorite things. I was in high school and was really into antiquing with my aunts. It has schedules of all the trains from numerous railroads. Somebody would replace these paper timetables every time they changed, and it's amazing to think this was a manual process until GTFS.

After college I worked in the KCMO City Planning and Development Department when discussions around light rail began.  It was a natural move to Portland as it's known for its transportation system and urban planning. That was one of the things that drew me in.

‍

How did GTFS come about? Can you give us the play-by-play?

It was actually Google Maps that sparked it. This was in the spring of 2005, and Google Maps had just come out with driving directions. I had come back from San Francisco, where it was really difficult to get around on public transit because of the different agencies. I was working late in my office, and it hit me that it should be just as easy to get transit directions in Google Maps as it was to get driving directions.

At that moment, I knew this was really big and was going to change things. I was wondering, are other people going to realize how big this is?

I ran down the hall to see if Carolyn Young [former Executive Director of Communications and Technology] was still there working late. She was, and I told her about it. She said, “That’s brilliant, go do it and let me know if you need anything.” And I did need to call on her a few times.

It took me a few months to reach out to Google. I had tried everything, and finally I sent a message out to a connection who had a friend at Google, Chris Harrelson. Chris had been working on adding transit directions in his 20% time. It was serendipitous that we were connected at that time.

‍

When did you get to a point where you realized, we now have a format? The urban legend is that GTFS strongly mimics the database structure of how TriMet stored its schedules.

Yes, it does.

The next step was to get Tim McHugh [then Manager of Enterprise Systems Development at TriMet] in the room with Guy Tinat and Mike Gilligan, two engineers. Tim was the one who designed TriMet’s database.

Transit data is spatial and temporal and can be complicated to newbies. Chris had really done his homework. They were impressed.

Tim did a CSV dump of our tables and sent it to Chris. They had started working on it. We had to sign NDAs, the four of us and Carolyn Young. No one really knew about it, not even my dog.

I really want to give credit to Tim. This is his brains. It wouldn’t have happened without his database design. GTFS still resembles the original format he provided it in.

‍

So you were working on it through the summer. When did GTFS launch?

I knew how important a standard was. I invited Chris and Joe (Hughes) Trellick, the two Google engineers working on it, to Portland along with several other transit agencies. We had a two-day workshop.

The goal was to get other agencies on board and put their data in this format. After a day, I felt it was not going to happen. One of my colleagues at King County Metro said to me on the side, “Google is going to have to pay for my data.” I felt like it might fail, but it just took time.

It was in December of 2005. I couldn’t know in advance when it was going to launch. Chris called me late at night and he said, “It just launched, and you won’t believe the numbers.” It happened very fast.

‍

How was it on the night of the launch?

It was exciting. It was great to hear the numbers. All of them were up all night just watching the numbers grow as the world woke up. I sent champagne to their office a day or two later. Coincidentally, there was an agency-wide leadership team meeting with the GM the morning of the launch and I rushed to share the news with Carolyn so she could make the big announcement.

I was glad that I always kept a suit jacket in my office. There were TV stations coming in all day long, radio stations wanting interviews. It was nonstop.

I had to buy a few suits and shades of lipstick. The PR department gave me tips. It was really exciting. That was during the week. My colleague who told me Google would have to pay for data called me Sunday night. He said, “I have a meeting with the General Manager tomorrow at 8am. He wants King County Metro to be next.”

It just made sense. What was happening at the time was there were a lot of developers who were interested in getting into [transit data]. They were screen scraping the data off of websites. We were the first agency to actually make our data openly available: developer.trimet.org. A lot of developers started using that. They wanted to do it worldwide, in New York and places like that.

The idea of opening your data and making it publicly available was just not mainstream at that time.

‍

I remember the same time period. I first heard about Google Transit in December of 2005. It got a lot of attention and press. At that point, did you think, “This just revolutionized transit data forever?” Did you realize how important this was going to be?

I knew that when I first thought of it. I’ll never forget that moment at my desk. Again, I knew how big it was going to be. But I wasn’t sure others would. Carolyn Young was a visionary, and I’m very grateful to have worked under her. At the time she knew how important it was and supported it. Even at TriMet, when it went live there were grumblings. That’s when she went in and had to make some phone calls.

‍

If you think back to the following years when every agency was thinking about whether to adopt the GTFS format or not, eventually it became a “when,” not an “if.” What do you think the motivating factor was? Was it riders demanding that, or did agencies already want to provide GTFS data?

At first, GTFS was called the Google Transit Feed Specification. Google quickly realized it was in their best interest that it was widely adopted. We were at dinner and they agreed to change it to the General Transit Feed Specification. We realized that was one of the keys to adoption.

I did a lot of advocating. I traveled a lot. Other agencies brought me out. New York, Israel, Vienna, Paris—I had a lot of people wanting me to come talk. There were a lot of other people within agencies that were trying to raise the visibility of the GTFS standard. It didn’t happen by itself.

I think wide adoption could have happened even earlier. It was challenging from a technical perspective for a lot of the agencies because they did not have centralized databases. There were a lot of siloes.

The core of that data comes from the scheduling system. Hastus and Trapeze at that time had developed an export for GTFS. That facilitated agencies adopting the standard.

Several companies wrote papers with regards to the inefficiencies and deficiencies of GTFS. “CSV?” But we wanted to make it as simple as possible for agencies to adopt. So that helped facilitate it. I think TriMet was in the best position to make this go fast, to make it happen in six to nine months.

I’m wearing a t-shirt that Google produced for an APTA conference in 2007 right now. It’s pretty rare; there’s not too many of them left. I even have a tattoo in commemoration of GTFS, OTP, and shared use mobility. I did each of those on the anniversary of the standards.

‍

It’s now almost 20 years since that launch. There are still cases where agencies may buy a new technology that uses a proprietary interface and does not support GTFS. If you were to give advice to agencies about how they think about new technologies, what do you think the importance of open data standards are in that context?

You put that in your contract. It should be standard in any contract.

Open data, open architecture—especially with so many working technologies out there. You have to have standards for it to work, for interoperability. You have to have it in your contract. And you have to own the data.

‍

You worked on GTFS and GTFS-rt. You and I worked together on GTFS-Flex. Swiftly has been working with Transit on evolving the standard to now handle detours and situations where there’s a gap between the schedule and real-time, where things are happening that are atypical. Are there areas like that where you think GTFS should evolve?

The detour spec, Trip-Modifications—I think that’s just really smart. It’s a really good example of the beginning and evolution of a standard.

Learning from GTFS, the benefit to the customer has to be visible. Google Maps at the time was big enough to get the attention of transit general managers and the public.

I think especially after COVID, there’s an influx of needs for real-time information, especially with regards to safety and security. A lot of the foundational systems have untapped information that could benefit the public. Agencies right now are focused on getting ridership back, improving mobility throughout the region, making life more livable. More accurate real-time customer information is one of the keys to doing that.

‍

One of the things that excites me about the GTFS Trip-Modifications standard is that the data format for detours is the same evolution of what it would take to do shuttles, bus bridges, and more. That’s a huge gap in the type of information agencies give to their riders. When a train has broken down or rail service is not running and there are buses that are providing shuttle service, right now there’s very little information and trip modifications could evolve to cover that use case as well.

Yes, exactly. Especially in bad weather. You don’t want customers standing out at stops in the snow and the freezing cold. It’s imperative that they get to work. Trip-Modifications is going to be really important, especially for older adults and people with disabilities.

And it's great that we now have an organization like MobilityData to help coordinate the evolution of the GTFS specification to meet these use cases. That didn’t exist in the early days.

‍

Is there anything we didn’t discuss that you’d like to mention?

I remember traveling to DC and I walked into a room to present on the importance of GTFS adoption.The audience was full of black suits, white shirts and red or blue ties. At times it was very intimidating. I tried to make myself look older, to dress and mimic these men. I wanted to be taken seriously. Things have changed for women in technology and innovation over the years, but not as much as I had hoped it would.

‍

You’re a pioneer in more ways than one. Thank you so much for your time today!