Stichting Hartekind researches congenital heart disease in children. By using advanced data science techniques we helped them to cluster patients and biomarkers. For this we used an unsupervised learning model that shows which characteristics are comparable in these clusters. To improve the prediction of clinical events we developed a feature selection model. We identified that the recovery phase could provide valuable information for this prediction, something the researchers hadn’t looked at before. And we developed a tool that automates the interpretation of the numerous cardio-pulmonary exercise tests (CPETs). Stichting Hartekind welcomed our user-friendly model with open arms. 


  • three models in one week
  • helping very valuable research
  • doing good while having fun together

Curious what we can do for your organisation?

Would you like to know more this case or what the Amsterdam Data Collective can do for you? Get in touch with Dennis Diederix at, or check our contactpage.