On the one hand, everyone agrees that it would be good for the climate and the air in our cities if more people travel by bus and train instead of using their own cars. On the other hand, however, many car users argue that switching to the public transport system is not worthwhile for them because the prices are too high, the connections too bad or the means of transport too unreliable.

Modern data analysis could help transport operators to solve these problems and create better services. “However, the data collected in public transport comes from very different sources and data quality is often rather poor,” explains Prof. Dr.-Ing. Thomas Schlegel, head of the Institute of Ubiquitous Mobility Systems at Karlsruhe University of Applied Sciences. “This is why there have been hardly any approaches so far to use Big Data or Smart Data for reliability”.

Metadata are supposed to tame the data chaos

The declared goal of Schlegel and his colleagues in the VSB-ÖP project (“Reliability of Smart and Big Data in Public Transport”) is therefore to improve data quality. As there is often little to be done with the actual data and the different sources, they rely on metadata: In these, origin and typical sources of error are recorded, so that the basic data can be automatically classified and evaluated despite all differences.

The project team mainly focuses on the internal data from computer-aided operations control systems (ITCS for “Intermodal Transport Control System”): The vehicles provide, for instance, position data and information on arrivals and departures at stops, trains report the rotation of the wheels, and in some vehicles the number of passengers boarding and leaving the train is also counted.

All these recording systems are more or less accurate and are prone to errors at various levels. A typical source of error is the human being, for example if the driver forgets to log on to the system when leaving the depot. Data can also be lost during transmission to the control centre or arrive with a delay. And with all this information, one common question arises: Have the numbers been counted accurately or only estimated, are the data calculated or measured by a sensor?

The different sources and production methods lead to a jumble of data that transport companies have hardly evaluated so far - or only with classical statistical methods, which are of limited use here. “You could get so much more out of the data,” says Schlegel. His team is therefore working on visualization methods that will later be integrated into existing software. According to the IT scientist, big data and artificial intelligence are actually just a form of statistics - but one that can be used to find information that would otherwise remain hidden in the data chaos.

Schlegel cites a vivid example: a disturbance on the track. “As a rule, the transport companies then deploy their experts for emergency planning - people who have often been doing the job for decades and know exactly what alternative options the network offers,” he explains. But with an appropriate IT solution that can quickly calculate which routes and which vehicles can be used as replacements from the available data, it would be possible to react much faster to disruptions - and at the same time provide users with reliable information on how best to get to their destination.

So far hardly any funding for research on public transport

However, although public transport is more topical than ever, it has been treated rather shabbily by politicians, the public and even the transport companies themselves. Schlegel says that thematic funding in the transport sector is almost exclusively available for car traffic. In research on public transport, but also on cycling and walking, it is difficult to raise the necessary funds. The reasons are manifold, the scientist says: “On the one hand, there is generally little education on public transport in Germany. And for another, the level of suffering does not seem high enough to modernise a system from the ground up. The local authorities, which usually operate the transport networks, must already be interested in innovative approaches in order to adopt methods such as ours”.

Karlsruhe University of Applied Sciences has at least one reliable economic partner for this and other projects: the INIT GmbH, which has been providing IT solutions for public transport for over three decades, is also actively participating in the VSB-ÖP project and intends to incorporate the planned analyses and visualizations into its software. INIT customers will thus benefit directly from the project and its results. “But of course, we are also looking forward to further use cases that lie outside the project”, explains Thomas Schlegel. “All transportation companies that are interested in such solutions and want to provide their data for the project are welcome to contact us”.

As things stand at present, VSB-ÖP is going on until summer 2021, but the topic of Big Data, Smart Data and AI in public transport will of course not be put off after that. The team of Karlsruhe University of Applied Sciences and the INIT-GmbH is already planning the first follow-up projects - so that in the end there will be better public transport for everyone and the change to the public transport system will be worthwhile after all.