DB Regio’s line agent app (“DB Streckenagent”) has been available throughout Germany since 2017. If there are incidents, it helps customers to find out about the current situation on the routes they use and proposes possible travel alternatives in the event of disruptions. The app is optimised for regional and local transport and is therefore particularly suitable for commuters. But even long-distance lines can be monitored.
The line agent app was programmed by Siemens subsidiary HaCon in Hanover. The data that HaCon uses is made available to the company from the European Timetable Center of DB Sales and the traveller information system platform. The application compares the timetables with incoming real-time messages as well as additional information on cancellations and alternatives and generates the route information subscribed to by the end customer as push notifications. These are messages displayed on passengers’ smartphone without the user having to open the associated app, saving them the trouble of calling up information.
But the line agent app offers major benefits not just for Deutsche Bahn customers. Zero.One.Data (ZOD), DB Systel’s in-house big data start-up, has now tapped into the data pool of the DB line agent app on behalf of DB Regio as an excellent way of performing passenger analysis. ZOD sourced the necessary data from an external service provider and stored it in a database. Processing and visualisation is handled using open source tools based on the Python programming language, supplemented by information from OpenStreetMap.
Answers from the data lake
The messages on incidents and delays generated by the line agent app are stored in the DB Enterprise Cloud. 13 million data records are already stored there in the form of text files structured using the JSON data exchange format. They contain information on news, trains and subscriptions. Each business day, the volume grows by 200,000 additional records and one million push notifications. Deutsche Bahn customers have no need to worry about data privacy because no personal data is used.
“This data lake is ideal for building a database for error analysis,” explains Dr. Enno Middelberg, Data Analyst at Zero.One.Data. One particularly pressing issue for passengers is push notifications arriving too late. Zero.One.Data helps to identify the problem more clearly by means of data analysis and makes the data obtained available to the participating service providers for further error analysis. This enables them to determine when the late notifications were generated in the traveller information information system, when they were provided for collection there, and when HaCon called them up on the servers and finally sent them to passengers. The source of the problem can be defined only by investigating a vast amount of data. And this is possible only with big data analysis.
However, the analysis project commissioned by DB Regio revealed that the data lake fed by the line agent app holds even more information: “After a year’s experience with the app, the question of who was actually using it came up. And we had no answer to that,” recalls Daniel Preußer, Product Manager for digital product innovation at DB Regio AG and product owner of the line agent app. It was clear that commuters mainly use the app. “But we didn’t know where in Germany it is used and which notifications the commuter receives,” says Daniel Preußer.
We examined the data on commuters’ point of departure and place of arrival and then mapped the results to the postal codes.
Using statistical evaluation of points of departure and departure times, ZOD has closed this knowledge gap. Although data from the line agent app has been stored only since the beginning of the year, it already provides a wealth of information. Zero.One.Data’s analysis of the data revealed clear patterns. “We examined the data on commuters’ point of departure and place of arrival and then mapped the results to the postal codes,” explains Enno Middelberg.
The patterns were visualised by Enno Middelberg in the form of maps of Germany. Their level of detail is impressive: “We can now see how many times a week a commuter receives a notification about a delay and how often they receive an incident notification. We can also recognise where the hotspots in the notifications are located on the railway network,” explains Middelberg.
Big data-based mobility assistant?
Zero.One.Data was also able to provide DB Regio with an accurate image of its customers’ movements and the stations they use. The data showed that in the mornings, when commuters make their way to the population centres, the notifications are spread across a wide area. In the evenings, by contrast, the push notifications are mainly received at stations in the big cities. Middelberg explains: “The line agent app gives us information about the stations used by the customers and their coordinates.” “This helps DB Regio to better understand and characterise their customers and to provide even more personalised and customer-specific services and passenger information. Of course, we can also focus more firmly on eliminating the most common causes of incidents.”
Our vision focuses on the app as passengers’ ‘digital mobility assistant’
Despite the current issues with delivering push notifications, Daniel Preußer is sure the app will continue to develop positively. Additional features and product enhancements are in the pipeline for 2018. “Our vision focuses on the app as passengers’ ‘digital mobility assistant’,” says Preußer. After all, why not keep track of the entire public transport system, including all buses and underground railways? “The customer should be able to ask us questions and give us feedback,” adds Preußer. The goal, he says, is a personal mobility assistant that will make communication with passengers a two-way process – benefiting Deutsche Bahn and its customers alike.