Problem: Public transportation services often exhaust their capacity during peak hours. Beyond this, commuters often lack opportunities for last mile connections. Flexible service routing and individualization of trips are key challenges for future mobility systems.
Solution: The Geospin forecasting model accurately predicts spatial and temporal mobility demand for different modes of transport. The predictions are even possible for areas without current mobility service provision or without comparative data. With Geospin, you can optimize mobility network and fleet utilization, prepare for disruptions and create novel shared-mobility solutions (e.g. demand responsive transportation).
The model is trained using several hundred-thousands of mobility patterns from various cities.
Over the course of a day, mobility demand is being predicted and compared to the true values.
Peak hour demand can be predicted with 97% accuracy. Even for areas without prior connection to the system we can estimate the hidden mobility demand.
„Geospin provides strong technical expertise and powerful algorithms for various mobility services. The models proved to be particularly useful for transferring knowledge between different cities. That way we could forecast mobility demand for areas where we currently do not have any mobility data.“
Dr. Christian Schwingenschlögl, Head Mobility Data Analytics, Siemens Mobility GmbH
Dr. Christoph Gebele