Mobility Lab’s Transportation Techies brought tech lovers together last week at the Black Cat nightclub in Washington, D.C., to dig into the troves of data generated by the growing number of bikeshare companies in the capital and its surrounding communities.
Among other show-and-tells, Bikeshare Hack Night VIII presenters showed their development of bikeshare user profiles, their creation of new ways to map and check bike availability, and the ways they figured out how to make real-time data even more available to planners, researchers, and app developers.
Who uses bikeshare and how?
Joe Haaga dug into quarterly reports from Capital Bikeshare – CaBi, to fans of the docked bikes operator – for a thesis project aimed at identifying variables that affect ridership. Looking at data as diverse as ride time to the millisecond, origin/destination, and D.C. government maps, he graphed out how membership type, days of the week, and even weather influenced people’s behaviors.
As one might expect, ridership decreased on colder days, and casual riders spent longer on the bikes than members, who tended to use the system for commuting. Diving deeper allowed Haaga to break out the average number of trips per bike and create a heat map of highly utilized bikes. That showed bikes clustered around downtown D.C.
He created another map showing bikes that were rarely ridden, which tended to sit on the fringes of CaBi’s network in Maryland and Virginia. Such information about station usage is lacking and sorely missed for its ability to depict the actual reach of the region’s overall bikeshare network.
Representing Motivate, which administers CaBi as the D.C., region’s first bikeshare network, Daniel Gohlke (pictured at the top) introduced some new features on the CaBi app. For instance, users now see icons indicating how many bikes are in a given station. The icons showed up as hearts on Valentine’s Day.
For the smart-speaker owner, Christopher Williams taught Amazon’s Alexa how to tell users how many bikes are docked in any CaBi station, as well as how many spaces are available for parking. The information comes directly from CaBi’s General Bike Specification Feed and is supplied by asking, “Alexa, ask my local bikeshare the status of … .”
Alexa users currently need to fill in the blank with a five-digit station ID, but Williams foresees a “favorite stations” feature to simplify queries. Another tweak would provide predictions for a station being stocked with bikes at a particular time of day.
How to use bikeshare better
Kalimar Maia looked into when riding a Jump e-bike would be faster than driving. Video game-style simulations of cars racing Jump-ers indicated that only 10 percent of bikeshare riders would beat car drivers.
Undeterred, Maia pointed out that the Jump examples included confounders like the time it took to unlock the bikes but the car examples did not factor in the time it took to find parking.
And, of course, it would be interesting to contrast the cost of biking with the cost for driving in D.C., where parking often requires paying more than $10 per day.
For those committed to bicycling, Daniel Schep’s DC Bike Finder displays the locations of dockless bikeshare units throughout the city. The new universal app is not limited to Android and iOS devices, working in web browsers and on devices that use alternate operating systems.
Alexandra Ulsh and Michael Schade also worked out methods for locating docked and dockless bikes by creating heat maps for shareable bike clusters around the city. Schade’s map was even interactive, allowing users to include or exclude certain companies.
For her part, Ulsh wowed the crowd with a feature that shows how people use Jump bikes to ride uphill after work.
Discovering the data to mine
Tom Lee reverse-engineered bikeshare operators’ to “look under the hood” of what the companies communicate to users. Though he didn’t play with the data he unlocked, Lee did highlight the possibilities of doing so.
One result could be something very much like the API Report Card that Schep created for evaluating the quality of each bikeshare company’s data feeds and how easy they are to work with. That project left Schep so happy with Jump’s feed – and its bikes – that he wrote an alert applet called Jump Start for his phone to notify him at 7 a.m. if a Jump bike was within half a mile of his house.
Mobility Lab’s Jenna Fortunati helped with the reporting for this article.
Photos by M.V. Jantzen.