Veikkaus a forerunner in consumer understanding

The last Helsinki Data Science Meetup of 2017 was held at Veikkaus,  the Finnish state-owned gaming company that as of January 2017 consists of Fintoto, RAY, and (old) Veikkaus. Veikkaus operates all gambling games that are offered in Finland.

Of all gaming within the new Veikkaus, about 30% are currently in digital platforms and the proportion identified gaming is also constantly increasing. This of course opens up a lot of possibilities for customer analytics, and allows Veikkaus to follow more closely the problematic players – responsible gaming is a topic taken seriously.

After short introductions by Eetu Paloheimo on the new Veikkaus and the the business they are in, and Tero Kallioniemi about customer analytics utilized, it was time for data scientists and developers to take the stage.

Jarno Kiviaho talked about recommendation algorithms utilized when recommending the next games to be played by the user. He talked about the use of contextual bandits, the exploration/exploitation dilemma, and for instance gave valuable insights into how the system can handle cold starts.

Carlo Hyvönen introduced the recommender system API they have worked on, and the optimization of the architecture for scaling it up.

Finally, Andres Levitski talked about responsible gaming and how they are able to identify users at risk.

Interesting talks were complemented by delicious food, and as it was just a day after the 100th anniversary of the Finnish independence, we were served with delicious cakes on that theme.