Apps Are Not the Answer

It’s clear that apps, specifically mobile apps, are the most popular creations for civic hackers and are the usual output of countless hackathons nationwide.

But at the latest Transportation Techies, held for the first time in Baltimore, Justin Elszasz, an open data analyzer who blogs at The Training Set, said the focus on apps isn’t actually making transit systems better as a whole.

“You can focus on gadgets and tech all you want, but it’s not going to fix the system,” said Elszasz, who was one of several presenters to discuss their cutting-edge (and reader beware, at times a little more technical than the usual Techies Meetup) transit-related projects.

Justin Elszasz

Justin Elszasz

He noted that open data has been used to create apps like SpotAgent, which help people avoid parking tickets – helpful for individuals, yes, but not exactly addressing a systemic issue.

On the negative end of the spectrum, apps that help users avoid bad neighborhoods spread stereotypes, increase segregation, and – in many cases – are flat out racist.

600_437920342-290x145-1432838275Overall, if these apps aren’t responding to a civic need – like, for example, not creating higher-quality transit systems for the poor – then what are they actually doing?

Instead, Elszasz encouraged hackers to look beyond the app and focus on data analysis that can be presented to cities and communities and support better policy formation. He also suggested that hackers at hackathons should behave more like investigative journalists: spend a Saturday finding a story in the data, point out discrepancies in service or quality, and then present it to the public.

“It’s commendable to want to build something,” he said of the hackathons, but ultimately, to improve public transportation, sometimes an app is just not the right tool.

Here’s a look at a few other highlights from the event, which was included as part of the annual Association for Commuter Transportation international conference.

Reverse Engineering to Save Baltimore Big Money

Chris Whong

Chris Whong

For Chris Whong, a self-described mapmaker and data junkie, reverse engineering Baltimore’s My Bus Tracker program started with a Facebook post from the Maryland Transit Administration that read:

“…it would cost approximately $600,000 more to be able to format the data from our 25-yr-old CAD/AVL [bus tracker program] system into GTFS for use by outside developers [ie: to create apps].”

He, along with several other commenters, weren’t having it. By going to the Bus Tracker site, he discovered another problem with Bus Tracker: the user interface was bad. Really bad. And the company that built it, Trapeze, wasn’t just running a bus tracker for Baltimore, but for several other cities – all equally bad!

Being Baltimore based, he set out to create a better tracking interface by reverse engineering what he could find on the website. He was able to find the JavaScript Object Notification (JSON) and then run a POST (an HTTP method) to see what the code said for certain commands.

Once he had the basic code from the MTA, he then added the scraped data into GTFS and routes information. Next he added an additional JSON (/vehicles) and packaged it all using the open-source client Leaflet which specializes in mobile-friendly interactive maps.

Whong’s project attracted the attention of Technical.ly, which then attracted the attention of Transit App, which in turn wanted to use his API in their bus-tracking app for Baltimore. The project’s success was heralded as saving “Baltimore $600,000 in one day.”

While MTA was pleased with the app, they did dispute how much money had been saved and noted that the data presented in the app wasn’t accessible via email, phone or text-messaging. Still, the app was leaps and bounds better than what was previously available.

While this particular project is no longer in operation, Whong said it brought up an important policy question: is providing high quality data enough for cities or should they be creating apps?

Robots and People and Sidewalks

Kotaro Haro, a computer science researcher at the University of Maryland, wanted to find out one thing: how many sidewalks in Baltimore are usable for people in wheelchairs? The question turned out to be deceptively complex.

The most basic way to determine this was to perform physical street audits using Google Street View. Given the sheer volume of streets to survey, Haro crowdsourced the labor to Amazon Mechanical Turk. The instructions were simple: individuals would volunteer to mark the curbs where they found them and also to point out potential problem areas. Accurate, yes, but far too labor intensive to be practical in the long term. And, even though the Mechanical Turks are only paid a few cents per photo, the cost adds up.

Haro then turned to automation, creating a program by which a computer would recognize ramps. Much faster, much less labor intensive, but it was plagued by accuracy problems.

To combine the speed of the computer with the accuracy of human verification, Haro created Tohme, “remote eye” in his native Japanese. Essentially the program takes the street view photos and sorts them into easy and difficult tasks. The easy goes to the computer, the hard to human discernment. So far, human-only accuracy is 86 percent and Tohme accuracy is 84 percent. With the 13 percent reduction in costs, it’s more than a fair trade off.

The final product still needs to be verified the old-fashioned way. Haro said he plans to set up a volunteer website for individuals interested in being verifiers for a few hours. In the end, the project will create a detailed to-do list of where Baltimore needs to make improvements so that residents who use wheelchairs will find the city as accessible as those on foot.

How the Maryland Transit Administration Is Improving Its Data Quality

Jaime McKay and Michael Walk

Jaime McKay and Michael Walk

As open data becomes more prevalent for public transportation systems, the next battle is going to be to improve data quality.

Michael Walk of the MTA happily announced that in the very near future, in addition to overall Baltimore transit-ridership data, his organization will begin to publish ridership per bus route, per station, and also on-time performance data.

These additional data points will open a wide range of opportunities for coders, app-developers, and data-heads to fine tune their understanding of how people move throughout the city. Analysis of this data will assist planners in knowing which stations are over- and under-used, how to spread out the demand, and also where additional services are needed.

Walk noted that MTA also has data generated by customers via Rate your Ride that is also available to the public. Matching complaints and praise to the raw figures of riders could also yield ways to make commutes more responsive to customer needs.

Similarly, Jaime McKay presented ways that Central Maryland Regional Transit is using a partnership with Google to increase the accuracy of information provided via their popular TRIP app.

Photos by M.V. Jantzen

Share this item

3 Comments or Mentions

2 Comment(s)

Paul Mackie

Thanks for the comment, Dylan. We corrected the text.

Paul

Reply

Mentions

This article has been mentioned in 1 other place(s).