Can Big Data Help You Plan Your Commute?
Modern public transit systems are elaborate networks with dozens of stations and scores of trains serving thousands of passengers daily. Anyone who has ever stood on a platform waiting for the next train understands the frustration of delays. But for commuters in Stockholm, Sweden, things may be looking up.
Developed by Mathematician Wilhelm Landerholm of Queue AB and Stockholm commuter train operator Stockholmståg, a new algorithm promises to forecast delays up to two hours in advance. Via a forthcoming smartphone app passengers will be able to plan their trips more accurately and seek alternate routes if serious delays are in the offing.
Problems in transit systems usually result from accidents or delays of a single train somewhere in the network. This sets up a ripple effect, causing delays for the trains behind and possibly leading to disruptions across the entire system, even after the initial disturbance is resolved. This happens on highways, too, as traffic jams may take hours to clear even after cars involved in an accident have been removed.
Enter big data. Cars on the highway suffer from two problems: there is no monitoring system for tracking their movements and they are operated independently. Commuter train systems, however, do not have these defects. In fact, modern networks have traffic control centers with computer systems keeping track of each train’s location at all times. Ten years ago, this mountain of data would have been unassailable, but with today’s faster machines and this new algorithm it is possible to make accurate predictions about the future state of the train network in a longer time window. It’s a bit like weather forecasting but for your commute.
The algorithm is proprietary, so I do not know the particulars, but it works something like this. At any moment, the traffic control center has a snapshot of the entire system. This includes the location and speed of each train along with any disruptions. Using these data, the algorithm dynamically simulates the effect of each disruption on the entire network. In the past this forecast might have been possible for a few minutes into the future, but an efficient algorithm and a fast computer can extrapolate this out for longer time periods. The traffic control center can then make adjustments to deal with the shocks–add extra trains along various routes, for example–and then the new data are fed back into the algorithm for recalculation. Potential bottlenecks may thus be prevented or at least mitigated.
Of course, there’s nothing special about Stockholm; this sort of system could be implemented in any transit network with real-time monitoring capabilities. An obvious extension for the future is to incorporate passenger data to better predict the number and lengths of trains that should be deployed along each route. If it turns out to be as effective as promised then expect to see it rolled out in your local subway.
Math: predicting delays and decreasing your frustration.
Article originally appeared HERE.