Employees look surprised when they walk in through the staff entrance at Rabobank's head office in Utrecht.
In the lobby you can see art on the wall throughout the year, or a photography exhibition. But once a year the Modelathon takes place here. The Analytics community within the bank literally builds a greenhouse in the head office where new ideas are sown.
Marc Heemskerk, head of Risk Analytics: “Two teams have been concentrating on their own challenges for a week. Team Crop Suitability uses existing data to investigate which crops thrive best on agricultural land in African countries. Team Balance Predictor uses data to predict your future balance. This can be useful for both our private and corporate customers. The teams are made up of Rabobank employees with a background in data and analytics. They will work with different data sources, look at them from many different angles and work on their own business propositions. ”
Rabobank takes the Modelathon seriously. The 16 team members are planned a week off from their normal work and can fully dedicate themselves to the Modelathon.
The first edition of the Modelathon in 2018 gave rise to the startup Fundr and according to Heemskerk the wish for this edition is to continue to work on the projects after the pitches of the teams: “We want the business side of the bank to be more aware of the value of data and analysis. At the same time, our data scientists can move closer to the bank’s product teams to be of value.”
” We want to be more data-driven within the bank. Data itself is already interesting, but it only becomes more valuable if you analyse it and get better insights you can use in your products and services. ”
On Friday both teams pitch their idea. The first team shows in an interactive demo how you can predict where in Africa you can grow certain crops. Using weather data from the past 30 years, among other things, they carefully analyse the opportunities for farmers in various African countries. This gives the bank an idea of how they can help farmers with their financing needs and what the chances are of a successful harvest.
Ivo Hendriks is a Credit Risk Modeller at Rabobank and part of this team: “We have delivered a proof of concept and now the organisation is going to see what can be done with this project. Who knows, the model may become reality in the near future”.
The second team presents their ideas on how to predict a future balance for specific customers. In their demo, they start with private customers, but it soon becomes clear that the way of thinking can also be used for the bank’s larger customers. The analysts’ calculation method goes much further than copying the past into a possible future, as the explain in their thinking process and way of working. Data Scientist Freerk Venhuizen of the team explains: “Instead of assuming that we could fit everything into one model, it soon became clear that we had to make several models in order to look back on the one hand and to be able to look ahead on the other”.
In the audience, colleagues from various departments listen to the explanations and discuss how such data analysis can help them in their daily work.
With the Modelathon, two worlds have certainly come closer together, and Rabobank is taking things one step further to become a data-driven bank.