I work on advancing tools for human-centred data analysis. In contrast to machine learning methods that aim for automation (e.g., self-driving cars), human-centred data analysis methods aim to empower human users. In order to be effective, such methods and their results must be simple and interpretable yet provide valid answers to critical analysis questions.

I am interested in all aspects of data analysis methods: learning-theoretic foundations, efficient algorithms, and concrete applications. For the latter, I particularly focus on supporting scientific discoveries. For example, I collaborate with materials science researchers to pursue the data-driven discovery of novel functional materials.