You’re stuck on the Orange Line in the Rosslyn tunnel for 10 minutes wondering what the holdup is. Or maybe you’re standing on a crowded Red Line platform, thinking if a train doesn’t come soon you’ll just order an Uber or walk.
A couple of Washington-area data wonks and transit advocates have a solution: MetroHero, an app aimed at creating a real-time map of Metro’s trains. Its creators say it’s a way to keep Metro honest and ease rider anxiety — using the agency’s own data. If Metro says its trains are running at eight-minute headways, the app can tell you if that’s really the case.
But there’s a problem. Metro’s data is at times faulty, unreliable and riddled with holes. And when it comes to data, a few imprecise numbers can spoil the whole set. What’s more, the app’s developers say their efforts to address the problems with Metro have gone unaddressed for months.
“Metro has done some good on its promise of making its data available to the public for third-party use, but it hasn’t followed through, and so far it has actually shown a certain unwillingness to make change in a positive way,” said James Pizzurro, 24, one of two software engineers behind the free smartphone app.
The other is Jennifer Hill, a 25-year-old George Washington University PhD student in computer science.
The app, which launched on Android in September and on an iPhone-compatible website in December, has demonstrated its potential. In December, when Metro said it would run eight-car Blue Line trains all day for Arlington National Cemetery’s “Wreaths Across America” event, its data showed that only 40 percent to 70 percent of Metro trains were running eight cars at any given time. From a rider’s perspective, fewer eight-car trains means bigger crowds on platforms and trains.
And if you followed the data during last month’s blizzard, you would have seen the gradual drawdown of trains before the system shut down at a rate of around eight per hour until, poof, at 12:33 a.m. Jan. 23, there appeared to be no more trains running.
Unlike other systems, Metro does not advertise third-party apps or have any preferred clients — but the agency says its data is open to any developer. The New York City subway system, meanwhile, hosts a catalogue of 86 iPhone apps on its website, many of which were created by independent developers.
Metro’s open data interface feeds wait times for trains and other relevant information to software developers for apps such as MetroHero and MetroMinder, a similar app that lets users see performance data for the system. It’s like having displays from every monitor on every Metro platform on your smartphone.
Pizzurro says the information fed to his app is at times incomplete or erroneous, making it difficult to trust without confirmation from often-reluctant Metro officials.
Amtrak, MARC and VRE all offer real-time maps of their trains. Metro says it is evaluating the possibility of creating its own real-time map, but it says third-party developers offer many such options already. In New York and Boston, developers have drawn from those systems’ data to create real-time maps.
But in Washington, developers say they have encountered pushback from Metro. Metro says it has had phone, email and face-to-face conversations with developers over the past three years and continues to explore how its data can be used more effectively.
But as his frustration mounted over unanswered questions, Pizzurro petitioned Metro’s Riders’ Advisory Council in January, asking members to request that Metro work with developers to provide more accurate data through its interface. The 21-member body had already asked Metro to explore partnerships with third-party developers in an October recommendation. Pizzurro says he is doing just what the agency’s planning director advised at a gathering of transit geeks in October.
That director, Shyam Kannan, said at the time that Metro encourages developers to identify gaps in its open data.
“It’s really helpful for us for the ‘geeks’ to get together and solve problems,” Kannan said then. “The private sector is much more nimble and agile when it comes to app development. . . . If the private sector can tell us where the warts are [in Metro’s data], we can fix them.”
Pizzurro says he has identified a litany of such issues. For one, he says, trains that are idling for more than 30 seconds automatically disappear from the data interface, creating a chain reaction of inaccurate wait times.
Metro, in an email to The Washington Post, said that the “disappearing trains” are not a glitch, however. The transit agency said it knows where its trains are at all times but doesn’t want to make that information public for security reasons.
“If a train is holding due to a sick customer, a train ahead or another disruption, the train will ‘disappear’ because the [platform displays] cannot predict movement after an extended holding period,” Metro spokesman Richard Jordan said. “There are some security concerns with making trains uniquely identifiable.”
There’s also the tendency of trains to drop off the map when approaching the end of the line, and for trains in one direction to disappear during single-tracking — another limitation of Metro’s data, Pizzurro says.
The agency did not respond to requests asking why.
In New York, Michael Prude, a 28-year-old software engineer at Conde Nast, partnered with software engineer Ted Mahoney, 29, to create a real-time map of eight New York subway lines. Their data is based on live train locations fed by unique “Trip IDs,” something Metro does not provide for its rail system — citing security reasons.
This frustrates Pizzurro, who wonders why Metro offers Trip IDs for tracking its buses but not its trains.
“No one is asking [the Washington Metropolitan Area Transit Authority] to predict how long a medical emergency holding up a train is going to last. If a train is holding, don’t delete any trace of that train ever having existed in the first place,” he said. “It stands to reason that if a train was three minutes away from its next station before it started holding, it is still at least three minutes away from that same station while holding, and we think that’s valuable information.”
He added: “We’re curious to know what these ‘security concerns’ are, considering Amtrak, MARC and VRE all make their internal train identifiers available to the general public — in many cases, in real time.”
Metro did not respond to a request that it outline its security concerns regarding its data. It also did not make available an engineer or other staff member who could speak to the complexities of its software interface.
Kevin Ortiz, a spokesman for New York’s Metropolitan Transportation Authority, said issuing unique IDs for its trains was initially a security concern for that system, too.
“It was, but it was reconciled,” he said.
So, why is a real-time map so important to riders anyway?
Chris Barnes, curator of the Twitter account @FixWMATA and a member of Metro’s Riders’ Advisory Council who has advocated for MetroHero, says it’s about transparency and adapting to the 21st-century demands of a transit system. Plus, unlike a lot of Metro’s problems, he said, “this is one of these things that can actually be fixed.”
Barnes said it would be a way for Metro to generate some goodwill with riders, who are frustrated with chronic service disruptions and what many say is the agency’s unwillingness to be forthcoming with customers. Riders want information they can rely on. Without a patch, he said, Metro’s data has to be taken with an asterisk.
“They’re itching to have that discussion with Metro, they just need them to come to the table,” Barnes said. “Everyone benefits when data that’s made public is accurate and it’s presented in a proper way. It’s such a great opportunity for Metro to gain a little more respect.”
Prude, who helped develop the app for the New York subway, says a program such as his can foster goodwill in a different way. It gives riders an understanding and an appreciation for the complexity of a system, so maybe they are more forgiving when it does not work as it is supposed to.
“It’s really, incredibly easy to complain about your train being late,” he said. “But thinking about why it’s late, and the amount of work involved in keeping it on time, will hopefully get people thinking about the complexity of the whole system. That, we hope, will lead to better, smarter funding for infrastructure, new tech innovation for riders and a slightly less bad commute.”
This week, Pizzurro received an invitation to speak at a Metro “Hack Night” this month, organized by Mobility Lab, an Arlington-based transit research and advocacy group. It will be the fourth such event for developers of Metrorail-related apps, and the first ever at Metro headquarters. Organizer Michael Schade directed a reporter to the event listing, which shows that Metro strategic planning adviser Michael Eichler will be in attendance along with chief executive Carey Anne Nadeau of data consulting firm Open Data Nation, to talk “about the future of WMATA’s own [data interface] and open data.”