Passengers’ mobile phone data helps predict the next train

In many U.S. subway systems, estimating the arrival times of incoming trains can be more an art than a science. Enter NextTrain, a new innovation from New York-based digital agency Densebrain that predicts the next train in real time by using mobile phone data from those on board. Densebrain is perhaps best known for its CityMates line of mobile apps, which serve as transit guides for commuters, tourists and urban dwellers in various U.S. cities. Now part of the NYCMate app serving New York City, NextTrain is essentially a passive crowdsourcing system that uses changes in the Base Station Identification patterns of riders’ phones to follow a train’s progress and predict its arrival at the next station. Specifically, because of the lack of reception in the subway system, riders’ phones are unable to connect to base stations, or cell phone towers, along the route. Using an algorithm created by Mobile Transit Solutions — who have partnered with Densebrain for the creation of the app — NextTrain monitors that data in participating users’ phones and automatically submits anonymous reports to a central server, which aggregates the information and predicts the train’s arrival time at the next station in real time. NYCMate users waiting on the platform can then simply select the appropriate station on the NYCMate subway map, which will display both officially scheduled arrival times and real-time NextTrain predictions. Although NextTrain uses crowdsourced data, contribution of that data requires no input from the user; it also doesn’t affect battery life in any significant way, Densebrain says. Accuracy, as one might expect, depends on the number of people using the free app, which is available for both iPhone and Android; some 10,000 are needed to make it reliable in Manhattan, the New York Times reports. App-minded entrepreneurs around the globe: One to partner with or emulate in other transit-dependent parts of the world? Spotted by: Katherine Noyes