A new urban travel planning tool called Replica, run by Sidewalk Labs, simulates transportation trends using anonymous data that is expected to be far more accurate than traditional analysis.
It’s no longer enough to simply know how many cars travel a highway or how many passengers crowd onto a train or bus. Transportation and regional planners say it helps to have a sense of where riders are coming from, where they’re headed, or even which bike paths are perceived as safest for women or children. This level of information remains elusive, if only because tracking it in such detail can raise ethical and privacy questions among residents and city leaders.
Metro, an elected regional government serving Portland, Ore., and 23 surrounding cities, is exploring a possible partnership with the platform Replica, run by Sidewalk Labs, a data-heavy next-generation urban design company owned by Alphabet, parent company of Google.
Replica uses de-identified movement data (a process used to prevent the identification of someone) from about 5 percent of the population's mobile devices to glean information about travel patterns, and then creates a simulated model of travel with information about traveler demographics and the reason for the trip. Replica officials stress the location data from the mobile devices is de-identified prior to using it, according to a blog post by Nick Bowden, CEO of Replica. (The company did not immediately reply to an interview request.)
“They take a lot of movement data and build models of how people travel based on that, and then apply those models to a synthetic population that closely, but not exactly, matches our region,” explained Eliot Rose, technology strategist at Metro, who spoke with Government Technology last month.
The formation of a "synthetic" population is largely a planning exercise where small bits of census demographic data are used to build a new data set which closely matches the full population. That data set is overlaid with the movement data to "replicate" how the full population is actually moving through the urban environment.
“Because they’re using a richer set of data to build their model, they’re able to produce more detailed analysis,” said Rose. By the end of the summer, Metro expects to be in a position to make a decision about whether or not to accept the Replica data.
In the next month, Metro plans to have all the criteria and data in place to begin testing the Replica platform. More than 30 pieces of criteria will be explored. The criteria largely fall into two categories: the accurracy of the data and its ability to maintain personal privacy. If Metro goes forward with the Replica agreement, the contract will run for a year at a cost of about $457,300, according to project documents.
Typically, transportation or planning departments conduct complicated surveys in an area to get a deep understanding of who is using the network and how. The surveys are time-consuming, expensive and generally lack the sort of frequency needed to accurately take into account significant movements in mobility modes like the rise in the use of ride-hailing, for example. Replica data is updated every quarter, providing a far more accurate view of how a transportation system is performing, or in flux.
The Replica platform was selected for a number of reasons, not the the least of which is the data is simulated and de-identified. “This is an approach — as best as we can tell — that protects privacy by design,” said Rose. “It’s simulated data. It’s not real data."
“We’re definitely going to be testing this closely,” he added. “But, our hope is that Replica is striking that sweet spot between obscuring real people's trips enough to completely protect privacy; while at the same time, giving us a level of detail in information that we don’t currently have access to.”
Last month, the Portland City Council passed the Data Privacy and Information Protection Principles, which sets guidelines for “protecting private and sensitive data” managed by the city or those working for the city. The city will next begin working with various departments and bureaus to begin implementing the principles.
Replica also takes a look at the entire spectrum of movement, including pedestrians, people on bikes, riders on mass transit, and of course, autos.
"One of the most interesting aspects of the Replica data was to be able to get some visibility on non-motorized trips – people on bicycles, pedestrian trips. Most places in the U.S. don’t have a lot of actual count information on that,” said Ron Achelpohl, Director of Transportation and Environment with the Mid-America Regional Council, (MARC) a planning organization serving the nine-county Kansas City, Mo. metropolitan area. MARC is also involved in a pilot project to explore the application of the Replica platform.
“It includes that additional simulated information on who’s traveling and why. And that stuff is really critical to understanding the complete picture of the transportation system,” said Rose.
“If we want to make biking feel safer for people, we can learn a lot by learning where are the places where women and older adults feel safe on the road — how they feel about biking in different areas. That helps us get to the issues that we want to understand,” he offered.
“Similarly, equity is a really big issue in our region. And we could look at barriers to transportation and access to destinations among communities of color and low-income communities,” said Rose.
Since Metro is a regional government organization, Replica data can be shared by its numerous coverage area cities as well as transit agencies like TriMet, the Oregon Department of Transportation, Portland Bureau of Transportation and others.
“Under the structure with Replica, we will be able to make some of the aggregate data from this available to those partners, at no additional cost,” said Rose. “And so that’s really important to us. Because we know from our experience, when everybody in our region is looking at the same data, and using it to make decisions together, we make better decisions.”