Västtrafik AB in Skövde, Sweden
Third-party data in DILAX Citisense - using APC Data to improve collaboration with operators
Västtrafik AB, is the second largest transport authority in Sweden, taking thousands of passengers in the southern region of Västra Götalands län to work, shopping or leisure activities every day. To keep track of how many people are boarding on and off day by day, Västtrafik uses Automatic Passenger Counting (APC) systems. They have also been using the analytics software DILAX Citisense since 2019, to analyze the generated APC data. At the start the software only processed data from the train and tram fleet. One year later the bus fleet data was added– the challenge: The bus fleets were equipped with varying APC systems, so different data formats had to be integrated into the software.
Västtrafik commissions different bus operators do drive for them in different areas of Västra Götalands län. The bus fleets are equipped with varying APC systems, that are different from the DILAX system used by Västtrafik. To standardize the APC data formats and facilitate data processing, Västtrafik asked DILAX to support the integration of third-party data from the buses’ APC systems into DILAX Citisense. The pilot project was done with one operator (Keolis) for the scope of 22 buses. Other bus operators are to follow.
We have talked to Christer Karlsson, project advisor for Västtrafik AB, about this interesting software development project and about the desired improvements in fleet management and data analytics that Västtrafik is seeing.
Christer, thank you for taking the time to talk to us about the third-party data integration project. Let’s start from the beginning: Why did you decide to integrate third-party data in the first place?
The BI Team, Erik Andersson and Maurizio Aira saw the need to consolidate the data coming both from our own and the bus operator’s hardware, so it could all be processed in DILAX Citisense. Over the years it has become evident that some of the smaller bus operators that we work with do not have the capacity for data analytics. So essentially, they have an APC system, but no-one is working with the data. Since we are a very data-conscious and data-driven company, it was a priority to us to enable data-based evaluation of our fleet performance.
Why were different APC systems in use, instead of choosing one common system?
Overseeing the implementation of APC systems over the whole network ourselves is a major undertaking. Also, the individual operators know best the technical requirements for their fleets and what system will fit their needs, so we decided that they should choose and install their own APC systems. We have achieved a very high coverage of APC systems within the network, which is a good thing, but it did result in several APC systems being used, which in turn meant that we were faced with the challenge of synchronizing the data.
With the support of the DILAX team we are now able to import third-party data into the analytics tool we use, DILAX Citisense. The data from all sources is processed in there together, so reports are generated from all data, not just our own hardware, which gives us a much deeper insight into fleet performance and ultimately passenger behavior.
What aspects of your business has the integration of the bus data in DILAX Citisense improved?
It has for sure improved the quality and validity of data. The bus operator is sending us their raw data, which we then import into DILAX Citisense. Before, it was unknow to us how data was processed, and if we discovered any discrepancy we had to revert to the operator for clarification, which could take some time and effort. Since data processing is entirely in our hands, we can always be sure about what has been done, how the processing was carried out. And it has significantly decreased the amount of work on both sides: the operator does not have to worry about data processing and analytics anymore, whereas we do not have to convert data from one format to another or check back with the operator if we see any discrepancies. It benefits all parties.
How are you working with this data, which specific reports do you always conduct?
We use standardized reports that are created in DILAX Citisense and then imported into our own database. We analyze occupancy of lines and trips, the number of boardings and alightings at certain stops. We do comparisons to previous time periods, e.g., the previous month or week. We monitor which stops have the most boardings or alightings over the course of a day or a week. There is a model that we built using artificial intelligence to predict occupancy, based on historical data. This is also accessible in our customer app, where it shows up as a three-tier model (low occupancy, medium occupancy, high occupancy).
In addition, we use the number of passengers per month for revenue sharing between our company and our different operators.
The third-party data integration has already proved extremely valuable to us and saved a lot of work and time. The project is still ongoing, since we have several bus operators to integrate still, but we are very happy with the support we have received from the DILAX team.
You are also combining APC data with other data sources to evaluate network and fleet performance, is that right? Can you share what you are working on?
We use a lot of different data sources and are always trying to improve data processing in general to learn more about passenger behavior. For example, we are currently building an origin-destination matrix. It is relevant to us to know what the most-used routes through the area are.
Plus, we are using the APC data for different types of special investigations, which we can come back to in the future when we have some additional experience.
About Västtrafik AB
Västtrafik AB has been in operation since 1999 and is the second largest transport authority in Sweden. The company is owned and financed by the Västra Götaland region in the west of Sweden. There, it operates 1,857 buses, 101 trains, 263 trams and 36 ferry boats. Daily vehicle miles amount to 47,500, equating 12 laps around the globe.