If just one percent of drivers from commuter-heavy neighborhoods stayed off the road during rush hour, traffic congestion for everyone else would drop up to 18 percent, according to a first-of-its-kind analysis of cellphone data.
The finding provides a convincing incentive for people in specific neighborhoods to take the bus, carpool or work from home, according to research leader Marta Gonzalez, a civil and environmental engineer at the Massachusetts Institute of Technology.
Simply asking people in general to take public transit or work remotely is ineffective, Gonzalez told NBC News. The study shows that it's best to convince a small number of drivers from a carefully selected segment of the population to stay off the road during rush hour.
Going forward, the team aims to validate the findings on traffic simulators and then work with city planners to carry out the one-percent plan — that is, convince a small number of specific people to stay home from time to time.
The breakthrough comes thanks to the team's access to cellphone data. Previously, studies like this relied on travel diary surveys, which limited the data to a tiny fraction of commuters and a single day. The cellphone data, by contrast, allowed the researchers to analyze trips taken by 100,000 people over the course of several weeks.
“That gives us very good statistics of the flows and, more importantly, if you see a congested street you can have a good estimate of where the drivers are coming from and that connection is what allowed us to design the strategy,” Gonzalez said.
She and colleagues analyzed three-weeks-worth of anonymous cellphone data from Boston and San Francisco to obtain information about drivers’ routes, traffic volume, and speed on those routes. They were also able to determine the drivers’ neighborhoods.
All of this data was combined with information on population density and the location and capacity of roads in the two cities. This allowed them to determine which neighborhoods are the largest sources of drivers on each road segment, and which roads these drivers use.
By cancelling one percent of trips by select drivers in a handful of neighborhoods in Boston would reduce traffic in the region by 18 percent. In San Francisco, the effect was a 14 percent reduction in congestion.
The data itself was made available to the researchers from mobile carriers, who collect the information for billing and planning purposes. For example, knowing where most of their users travel in a day helps carriers decide where to build more cellphone towers, Gonzalez explained.
Her team's use of the data was subject approval of a review board that assured the study did not invade the privacy of individual users. “We are not allowed to make a study that can follow individuals,” she noted.
Since the methodology of the study requires only access to anonymous cellphone data, information on population density and a road map, it can be easily replicated around the world, which the researchers are now doing with an aim to reduce traffic in fast-growing cities.
The findings are reported today in the journal Scientific Reports.