City Planners that Ride On Data

By Pragati Verma, contributor

When Corey Acri, a lawyer and avid cyclist, moved to Philadelphia from New York in 2013, he noticed that the “biking infrastructure was a little off.” When he mentioned this to Mark Headd, then chief data officer of the City of Philadelphia, Headd told Acri there was no open data on bike route choices in the city, adding, “Why don’t you start a project and do it?”

Acri took on the challenge, despite his limited experience with coding. That summer, he began working with coding enthusiasts at Code for Philly, and Headd introduced him to transportation planners at Delaware Valley Regional Planning Commission (DVRPC). “I began by researching similar projects in San Francisco, Austin, and Atlanta, and soon collaborated with Kathryn Killebrew and Lloyd Emelle to create a mobile phone app that collects rider data in hopes of improving regional bike routes,” Acri said.

An App Is Built

Named CyclePhilly, the app lets cyclists voluntarily record their bike trips to help local planners and agencies understand bicycle trends, routes, and trip purposes, so they can improve bicycle facilities and connect the region’s trail network. Bikers who download CyclePhilly simply hit the “start” button at the beginning of their ride and “finish” at the end, while indicating the purpose of their trip.

“We got huge amount of data from CyclePhilly, and our hope is that a new version will capture an even bigger sample that will be much more representative of cyclists’ behavior as a whole.”

—Gregory R. Krykewycz, associate director, multimodal planning, DVRPC

DVRPC began collecting the crowdsourced raw trip data from the app in May 2014. They cleaned and processed it before visualizing it on a map, making sure to obscure the trip-ends to protect users’ privacy. They also snapped it to the nearest road or trail segment using a special algorithm that calculated and compared total volumes by segment.

“We got a huge chunk of new data, and it was a really important piece of our travel-demand model,” said Gregory R. Krykewycz, associate director, multimodal planning, DVRPC. At last count, in 2016, DVRPC had mapped over 12,000 individual trips by 300 unique CyclePhilly users. The idea was to supplement DVRPC’s existing demographic data—”bicycle counts” and “data from census on how many people cycle to work,” he added. “But we had no data on people’s route choices, length of trips, or how their choice varied with the purpose of the trip.”

Pushing Insights Further

DVRPC didn’t stop at using insights from CyclePhilly data to build better bike routes, but shared the anonymized trip-by-trip GIS data publicly. Independent researchers used it to assess behavior of Philadelphia cyclists riding the wrong way on one-way segments. For example, the study concluded that 42 percent of trips included a wrong-way segment, but only 2.7 percent of travel distance was wrong-way. They also found commute trips had a higher chance of wrong-way riding, compared to leisure trips. By identifying the demographic and network factors influencing riders’ wrong-way decision-making, the data helped planners and engineers choose between various types of bike infrastructures.

That’s not all: The CyclePhilly team also helped set up a similar project for bikers in Dusseldorf, Germany. As for Acri, the former lawyer felt that he was onto something big and jumped full time into web design with his company, AG Strategic Design. Calling it a “life changing experience,” the self-proclaimed bike fanatic explained that he’s currently working on a project called Whyabike to help cyclists “find a bike that’s right for them, and provide them personalized point-to-point direction based on a rider’s confidence level via text message.”

Meanwhile, Krykewycz is getting ready for a more ambitious crowdsourcing exercise. “We got huge amount of data from CyclePhilly, and our hope is that a new version will capture an even bigger sample that will be much more representative of cyclists’ behavior as a whole,” he said. His vision is to passively derive insights from riders’ phone data, with their permission, instead of requiring them to actively track and share their trips.

Several entrepreneurs, Krykewycz noted, are already coming up with ideas to help DVRPC collect a bigger dataset and derive intelligent insights to build a better transport infrastructure. As Acri summed up, “CyclePhilly is taking on a life of its own in a really good and positive way,” and inspiring many more data-driven analytics and insights projects to improve city planning.