civity and its project partners pushing ahead with the xMND research project
Based on initial data from the xMND research project, civity along with Telefónica NEXT, MOTIONTAG and the Fraunhofer Institute organised a hackathon on 22 August. The research project, which is funded by the Federal Ministry of Transport and Digital Infrastructure (BMVI), aims to use mobile network data to facilitate and improve demand-oriented planning in public transport. The aim of the hackathon was to improve detection of the transport mode in mobile network data.
When it comes to evaluating measures for public infrastructure, goals for the public transport turnaround and comparing public transport in different cities, there is one parameter that is frequently discussed: the modal split. This parameter is the distribution of transport demand per means of transport and is classically determined on the basis of transport surveys and corresponding extrapolations.
The enormous complexity and high costs of surveys like these mean that they are carried out at most annually and that they only provide a snapshot in time. Highly accurate smartphone-based data and nationwide mobile network data could be used to continuously calculate a modal split. One working group at the hackathon devoted itself precisely to this topic. An algorithm was developed that allows the modal split to be extrapolated to districts and cities. Using smartphone-based data and mobile phone data collected in xMND, a proof of concept was implemented for the city of Munich.
Validation was carried out via a specific comparison of mobile network data and user tracks, in which the two data sets can be compared directly for individual survey participants. This method makes it possible to provide continuous modal split data for cities, thus making it possible, for instance, to continuously monitor transport policy goals in cities.
In other working groups at the hackathon, the participants worked on detailed technical questions. An interface was developed which continuously plays back spatial context data, e.g. on transport services, for the dynamically changing radio cells. Extrapolation of test data as well as pattern recognition of the choice of transport mode were also on the agenda.