Uber’s open data platform needs some work before government can get excited

It’s a good start, but more work is needed.

That was the first reaction of Boston Chief Information Officer Jascha Franklin-Hodge had when reviewing Uber Movement, the ride hailing company’s first attempt at an open data platform to help transportation planners and researchers. In an interview with StateScoop, Franklin-Hodge praised Uber for opening its data, but said richer data and greater accessibility is needed before Boston can benefit. 

“I think that they’ve kind of recognized one way in which the data that they’re gathering can be really valuable, but I think the biggest challenge that we have with it is first and foremost that they’re not actually making their data available,” Franklin-Hodge said.

Uber’s new website, launched last month, presents visitors with a map to compare trip times between different sections of cities and do comparisons to see travel time differences depending on the date, time of day, or based on an event like the Boston Marathon, during which time road closures are common. Franklin-Hodge said this is fantastic, but Uber’s data can only be downloaded for individual searches. There are also only a few cities where data is available — Boston; Washington, D.C.; Bogotá, Colombia; Manila, Philippines; and Sydney, Australia.

First impressions

Uber Movement is helpful when answering highly specific questions, like how long traffic delays might be for certain road closures or construction, but overall, Franklin-Hodge said its use requires a lot of manual searches and its insights are limited. The data cannot be downloaded in bulk, nor is the data provided in real time. The site also lacks an API, preventing outside developers from easily grabbing data for their projects.

Beyond such accessibility issues, Franklin-Hodge said what is most noticeable is an absence of high value information, like demand and ridership data. Demand data would help the city see how they can fill gaps with buses or rental bike options. These data sets are particularly noteworthy, he said, because if Boston’s transportation planners had demand data they might be more willing to dedicate curbsides for ride hailing services — something Boston doesn’t have today, but that Uber is seeking in cities around the world.

“While this tool is sort of one lens to look at the transportation system it really doesn’t do a whole lot yet to help us with that kind of longer-term, bigger-picture plan,” Franklin-Hodge said.

In Boston, like other densely populated cities, planners are trying to decipher a jigsaw puzzle of transportation options. Buses, bikes, ferries, subways are all taken into account along with personal vehicle traffic. The goal cities have, Franklin-Hodge said, is to combine these to make transportation as affordable and expeditious as possible. Without demand data, bulk downloads or an API, Uber Movement struggles to impress.

“We’re trying to encourage more use of public transportation, more use of active transportation methods [like walking or biking] and to really think about how to integrate Uber, Lyft and similar services into these broader transportation goals,” Franklin-Hodge said.

Further data from Uber could inform the relationship between usage and car ownership and how parking availability and pricing change with the greater introduction of ride hailing.

“I think ultimately, what we do know is that the [ride-hailing] services are a huge and growing part of our transportation system and if we’re doing our jobs we actually need to build the physical infrastructure, and the policy frameworks and all that stuff, to accommodate these services but we can’t do that without any insight into how people are using them and where and when,” he said.

A work in progress

Jordan Gilbertson, Uber Movement’s product manager and Andrew Salzberg, Uber’s head of Transportation Policy and Research, acknowledged the platform’s constraints and said it is still in development. More cities will be added soon, the company has other data sets it’s considering adding, and an immediate priority for the development team is to add an API and an option for bulk downloads.

Uber doesn’t see Movement as a one-off data project, the executives said, but a long-term commitment that will continue to add data sets and analytics tools.

“We see this very much as a starting point” Gilbertson said. “The entire idea of this is that we are building a product and that is in turn backed by engineering resources and that is backed by an entire team to build this out over time and work very closely with city officials and transportation planners.”

Yet when asked about demand data, Salzberg said it is unlikely that Uber — or any other ride hailing company — would be able to release such information publicly due to the competition in the market.

“I think the issue with demand data, and you can talk to other companies in this space, is that there is a really intense sort of competition right now,” Salzberg said. “Many companies are looking at specific times of day and specific locations and targeting marketing promotions or actions to ensure that drivers are in a certain area at a very fine grain level, which I believe wasn’t the case before services like ours was available.”

While demand data like ridership numbers might never be public, Salzberg said that Uber does have unique data sharing relationships with individual cities. These may or may not offer an opportunity to share some form of demand data with cities privately.

Even so, Gilbertson said Uber is considering for release of safety data and other information that might inform infrastructure decision making.

“We’re also interested in how we can start to bring in totally different sources of data that are also valuable, things around safety, infrastructure or any other type of thing that has to do with the state roads and the state of safety on those roads.” Gilbertson said. “You know this is how people move around. It’s leveraging a phone to understand a city and the way people move throughout the city.”