"Don't be afraid to start, because it really pays off."
Publiek Vervoer Groningen Drenthe is an organization whose aim is to keep public transport accessible, while adapting for future trends such as the ageing population in those regions. They offer different modes of transport to specific target groups (such as the elderly and students) to ensure efficient travel and an inclusive society. These target groups get access to ‘WMO taxis’ for a reduced price.
Desired insights
With transport costs projected to rise in addition to the impact of evolving population characteristics, Publiek Vervoer wanted to explore whether they could reduce the number of people that make use of WMO taxis, by promoting the use of public transport instead. This would help them better distribute the supply and demand of mobility services, creating an overall more efficient, and less car-reliant transport ecosystem.
To facilitate this, data was needed to better understand user behavior and transport patterns. “In essence what we needed to know is, which travellers can we get to use public transport, and how, and for who are the WMO taxis an absolute necessity”, says Petra Buitenhuis (Director a.i. of Publiek Vervoer). A pilot project was initiated to collect data. Their target audiences were provided free public transport passes alongside their WMO passes, to see if more people would choose public transport instead of a taxi to get them from A to B, now that it was offered free of charge.
Why Roseman Labs
To get these insights, Publiek Vervoer needed to compare data from multiple sources. “The challenge we are facing is how to compare data with each other in an AVG-friendly way”, says Buitenhuis. The public transport network in the Netherlands is comprised of many separate organisations that each have their own data sets. This was problematic because:
- Travel data (travel history and user demographics) falls under GDPR, making it confidential and only shareable with permission.
- Travel data needed by Publiek Vervoer is stored in multiple locations, across multiple providers.
- Travel data within Publiek Vervoer organizations may not all be recorded or structured in the same way.
The solution
Roseman Labs’ Virtual Data Lake addressed the confidential nature of the data, the privacy commitments of all the participants and the challenges of dispersed, non-uniform data. Data sets from each contributor were encrypted at source and then subsequently fragmented to be spread across three separate servers – each holding a ‘piece of the puzzle’. Before any analysis was approved or took place, the Roseman Labs team performed ‘blind’ quality checks to verify the formatting of data provided (e.g. date of birth notation DD-MM-YYYY vs YYYY-MM-DD). Although it helps to have data sources that use the same standards, it’s not a necessary requirement for a project to start.
The results
“You can’t share insights on an individual level, such as “Mrs. Janssen, who lives on the Wilhelmina street travels by bus very often, regularly uses FLEX transportation and occasionally makes use of a share bike”, explains Petra, “but we did gather valuable insights on a higher level”.
