Benchmark data set Person utilisation of regional and long-distance railway stations in Germany

This data set serves as a benchmark for predicting the utilisation of people at German long-distance and regional railway stations. The personal utilisation was determined on the basis of the displayed load on Google Maps based on floating phone data (data © 2021 Google). In addition to the actual utilisation rate, the dataset also contains the historically average utilisation rate per hour. The dataset is divided into a training, validation and test dataset.

The benchmark for predicting categorised utilisation is a mean-absolute error of 0.49 and an F1 score of 0.53 on the test set.