The demands on transportation systems continue to grow while the methods for analyzing and forecasting traffic conditions remain limited. Here the authors show a parameter-independent approach for an accurate description, identification and forecasting of spatio-temporal traffic patterns directly from data.
