And other news from Transportation Techies
Boulder B-Cycle, the city’s bikeshare system, is looking to expand its ridership base. While it has doubled the number of annual rides between 2014 and 2015, from 43,000 to 84,000, the Daily Camera reports the 300-bike, five-year-old system “is only operating at a fraction of its capacity.”
However, an important step in building this base is understanding who is already riding and why. During May’s special Boulder, Colorado, edition of Transportation Techies at Spark Boulder, a group of developers presented several visualization tools to better understand the trends underneath current B-Cycle ridership.
Pass types predict usage
Monish Prabhakar examined 13 variables within B-Cycle data to examine the nature of the system’s usage. His biggest discovery among these variables was how riders with different pass typesuse B-Cycle. Across the four pass types – annual members, monthly users, daily users, and single-trip users – there is a marked difference of where and when riders check out or return a bike.
Annual members ride the most on weekdays while daily and monthly pass holders were more frequent and dominant on weekends. Students who have a semester pass (a great idea for D.C.’s Capital Bikeshare, by the way) tend to check out bikes later in the evening. These differences are pronounced enough that Prabhakar has begun to develop a machine learning program, which uses past data to teach itself, that will predict the type of each rider’s pass based only on the trip duration.
Tyler Byers built upon the pass analysis to determine how far B-Cycle users ride using the shortest possible real route between their start and end points. Like almost all bikeshare systems, B-Cycle does not use GPS to track the actual routes, only the start and stop points. With this, Byers was able to visualize how riders likely move around Boulder, as well as estimate the average speed (7.5 miles per hour) and average trip duration, which matched Prabhakar’s analysis.
To close, Gareth Coville explored the possibility of identifying the differences among different purposes behind rides based on usage data. Like Byers, Coville created a map of the most likely route each bike took between check-out and check-in. Through these, he identified the most heavily traveled corridors of Boulder and, importantly, where ridership is lacking despite the presence of a bikeshare kiosk.
Coville also teased out the average distances and duration of each membership type, with annual or semester pass holders returning the bikes quickly and after relatively short distances. One- and seven-day pass holders rode for longer distances and periods of time.
Using this data, Coville hoped to determine if it’s possible to identify which rides were commuters and where they are riding. From the busiest kiosk, Coville found that many of these commuter rides begin at the transit center in the morning and end there in the afternoon, suggesting that many users actually come from out of town and use the bikeshare system to complete their trips. Understanding this dynamic will be important for B-Cycle to determine how to target locals to increase their usage of the system.
Parallels across modes in sharing systems
An interesting parallel arose between B-Cycle and car2go patterns, even across cities. Brian Timoney analyzed six months of car2go usage in neighboring Denver and found trends that turned out to be similar to how people use bikeshare systems.
Just as Boulder’s core is popular for B-Cycle users, Denver’s has been the hotspot for car2go members, who then take the cars for fairly short distances. In contrast, the fringes of either system’s service area is generally neglected by members, seeing significantly less usage than units towards the middle of the system. However, bikeshare systems like B-Cycle have the advantage of efficient rebalancing, while car2go crews can only redistribute one vehicle at a time.
While both B-Cycle systems in Boulder and Denver have shown consistent growth over its lifetime, Denver’s car2go has had to reduce its service area as a result of struggling numbers.
Take a hike
Also, if you plan on taking advantage of the Boulder region’s outdoor resources (and you should) – check out Trailsy. Trevor Ackerman presented this detailed app to help trail users identify available resources around the area’s trailheads. Pulling data from the City of Boulder, Trailhead Labs and Boulder Open Space and Mountain Parks, users can identify different amenities at each trail, as well as trail conditions, lengths and park-ranger tweets for up-to-the-minute information.
Data tools like those in these Techies presentations can be informative offering lessons to how bikeshare systems can best engage their ridership base and build to better serve it. Every city has a unique dynamic among its transportation users, and understanding specific behaviors or trends helps to establish an effective campaign to get more people onto shared bikes.
Photos, from top: Andy Gup speaks at the Boulder meetup (M.V. Jantzen, Flickr). See more photos from the event here.