Martijn Westra reports on a recent hackathon, which aimed to boost health outcomes and demonstrate the value of both data science and actuaries
On 1 July 2022, the Dutch Royal Actuarial Association organised its first hackathon. The goal was to show the added value that actuaries, with their broad technical skillset, can have in the data science domain – and also to make actuaries aware of the added value that data science can bring both within and outside of their domain. The 10 teams that joined the event were asked to come up with innovative and creative solutions to reduce obesity and allow people to live longer and healthier lives. Making use of data science techniques, the teams found some interesting results – and even had creative ideas for new products.
The hackathon
The teams were given a dataset of results from the Behavioral Risk Factor Surveillance System, which collects data on US residents’ health-related risk behaviours, chronic health conditions and use of preventive services. It consists of more than 400,000 respondents, who answer more than 300 questions. Besides demographic information, it contains data on factors such as health status, healthcare access, alcohol consumption, exercise and diet. The teams were also allowed to use any other relevant data source they could find to enrich this information – for example from the Dutch Central Bureau of Statistics.
The teams approached the problem in different ways. Some brought a machine learning toolkit in Python or R and went all out on the data with gradient boosting and random forests, trying to find relations between factors and causes of obesity. One team found that children who grow up in troubled families are, on average, more overweight. They proposed investing in organisations that help families overcome such problems. Other teams started putting Post-it notes on a piece of paper to define their strategy. This approach led to a more balanced use of (advanced) data science techniques and strategic decision making. For example, one team advised people to go and live on Texel (a small island off the Dutch coast) to give themselves the highest probability of a long, healthy and happy life.
At the end of the day, each team provided a pitch of one minute, after which the jury could ask some follow-up questions. The winning team struck a good balance between the use of data science techniques and the creation of an innovative product. It proposed an app that provided a personalised approach to improving lifestyle; using the provided dataset, it was able to plot the scores of a certain lifestyle component (for example the amount of vegetables in a diet or amount of time spent on sports). The app then provided advice on the steps the individual could take to improve their score.
Applicability
With the improvement of computational power and increased availability of data, advanced data science techniques have become popular, and have been widely used for several years – for example in self-driving cars and chatbots. Applications based on machine learning techniques have also been shown to outperform humans at playing games and recognising handwritten numbers.
“It is not just that data science adds value for actuaries – actuaries also add value to data science through their ability to understand, build and explain complex risk models”
These techniques may be useful in the field of insurance, for example in making claim decisions by assessing a picture of a damaged car. Looking at the work that is typically performed by actuaries, data science techniques may be useful for, say, improving the pricing of non-life products or the curve-fitting process of Solvency II internal models.
The hackathon showed that actuaries can understand and implement data science techniques with relatively little effort. However, it is not just that data science adds value for actuaries – actuaries also add value to data science through their ability to understand, build and explain complex risk models. Now it’s up to actuaries to find good use cases and add additional value within – and maybe even outside of – their field of expertise.
Martijn Westra works as a balance sheet manager at Athora Netherlands, and was part of the working group that organised the hackathon
Image credit | Shutterstock
