Big Data and Mass Transit: From Bikes to Buses

“I have travelled the length and breadth of this country and talked with the best people, and I can assure you that data processing is a fad that won’t last out the year.”

-- Editor, Prentiss Hall Books, 1957

Last week, the Globe profiled an effort by 23-year-old entrepreneur Matthew George to use data analytics to provide “pop-up” bus service across many underserved routes in the Cambridge-Boston area. This “pop-up” service—called Bridj—is designed to use data about “where people live, work, and play” to predict where non-stop service is needed and adjust schedules based on time of day/day of week, etc.

George’s introduction of disruptive analytics to the metro-Boston transit network is long overdue and I’m anxious to see how his system works (and to try it myself come July 4th weekend). But, as noted by MIT Professor Nigel Wilson, George’s service (which is expected to launch at a cost of $5-8 pre trip) has the potential to siphon riders from the MBTA. Indeed, while the Bridj homepage champions “Better Transit. For All,” it is not yet clear whether the business model can rely solely on routes not directly served by the T.

In a normal setting, competition would be an unquestionable good—with the better product/price/service winning out over time. However, public transit is a unique animal—a deeply subsidized public good that must cater to the needs of very low-income city dwellers (among others).

To his credit, George seems quite cognizant of this fact and has indicated that he hopes to reduce fares to a price approaching a single-ride T-pass ($2-2.50). However, it is ultimately not the job of entrepreneurs like George to worry about how their innovations might affect competitors like the MBTA.

Instead, as I briefly noted last year, what the MBTA and other transit agencies from New York City’s MTA to the smallest regional network in Berkshire County need to do, is to get in the data analytics game themselves. In Boston, this effort should include investing in smaller vans that can operate at lower cost than articulated buses, depending on demand, GPS tracking to allow riders to plan their trips, and demand-responsive transport during late nights and weekends. In the spirit of George’s “pop-up” service, demand responsive transport covers a fixed service area but without fixed routes, allowing it to cater to fluctuations in ridership.

This type of planning should not be limited to buses, but should instead be used to integrate a municipal transit network’s bicycles as well. In NYC, CitiBike recently released a trove of data charting hundreds of thousands of rides and, as shown in the graph below from the NYU Rudin Center for Transportation Policy and Management, there is a slight, but meaningful correlation between subway disruptions and use of CitiBike along those routes.

Dubbed “reactionary biking” by the Rudin Center, this pattern should lead to partnerships between the MTA and CitiBike. For instance, when there is a planned service outage—especially a long-term outage, like the 5-week closure of the G train’s Greenpoint Tube planned for this summer—MTA should not only provide replacement bus service, but also work with CitiBike to extend bike share to affected communities. Similarly, the two systems should share data on ridership so that CitiBike can do a better job of balancing stations near transit hubs which, at certain times of the day, are overrun with passengers (most notably on the Lexington Line (456)).

In 2012, Peter Sondergaard of the Gartner Group declared, “Information is the oil of the 21st century, and analytics is the combustion engine.”

If Sondergaard is right, public transit systems cannot sit back in the horse and buggy age while private companies like Bridj act like the Maseratis of the data world. They need to get in the game themselves and use “big data” to increase efficiency and improve service for the millions of Americans who rely on buses, trains, trams, and bike share.