“The first automated transit forecasting tool of its kind.”
“For commuters in Stockholm, Sweden, things may be looking up.”
The Train Brain is based on an algorithm that can predict when trains will be delayed, letting commuters avoid delays and better plan their travel.
Using years of historical data, this forecasting model can predict delays hours before they happen. By monitoring live arrival and departure times, The Train Brain knows when a train is late and forecasts the risk of further delays across the entire network.
The model works in a similar way to how a seismograph monitors earthquakes, it looks for significant peaks. In The Train Brain, these peaks are represented by trains pulling into a station later than the scheduled arrival time. The Train Brain automatically compares how a similar delay on the same line impacted the network on previous occasions to predict how it will affect the network in the near future. And much like a human brain it evolves. It learns and improves its ability to make forecasts over time.
“Don’t ask how accurate the forecast is at a specific point
in time, ask instead how fast the forecast is recalculated.”
Wilhelm Landerholm – Data Scientist & TRAIN BRAIN FOUNDER
They had a saying in ancient Rome: Man is not allowed to know what will happen tomorrow. This is of course true. Even though we can make very exact predictions with today's computing power, forecasts are by nature inexact.
So how accurate is The Train Brain? That may be the wrong question to ask. A train network is affected, every minute, by a multitude of events. Some are scheduled and some just happen randomly, like accidents or someone illegally running across the tracks. This means that even if a forecast is almost 100 % accurate when delivered, it may be obsolete only 10 minutes later.
One of the main benefits with this solution is that it completely recalculates its forecasts, for an entire traffic system, within one minute. So if the circumstances change the next forecast is only a minute away, including the ripple effects of this new event.