Identifying relations between Feyenoord Seasonticketholders based on access control data

Sports Alliance manages data for professional football clubs. From the prediction models we have developed and deployed over the years, we learned the importance of existing relations between fans. Renewing or buying a season ticket, it turns out, is quite often dependent on family relations. Family relations which we 'know of' through our deduplication and grouping processes. However, this narrows a relationship down to people sharing the same name and/or the same address.

We were convinced that other more casual relationships between season ticket holders, such as friends and neighbours, will also be an important factor in buying or renewing season tickets. However, short from asking supporters, there was no way to know if a relationship existed. So when we were invited to a data science student hackethon by the Erasmus University, it wasn't very difficult to find a client willing to participate. In fact Feyenoord was very keen to participate in the project.

One of the student teams decided to address the issue of how to identify relationships between supporters based on the access control data. The idea behind this question was that a 'groups of friends' would typically enter the ground through the same entrance and shortly after each other. These friends would then, probably, sit next to each other or at least in the vicinity of the other. As it turns out, it is possible to identify relations between supporters.

This presentation will explain in detail, from data preparation through the use of statistical techniques in R, how these relationships can be identified.

This lecture is made possible by Tableau.