Data are not neutral facts, but rather relational objects whose meaning can only be discerned in relation to a specific context. This “relational view” of data means that the values of those who collect, process and analyse data affect what counts as data in the first place, and how data are interpreted. This has big implications: fostering the responsible use of data as evidence is crucial to produce reliable empirical knowledge, and needs to be fostered and rewarded.

My work over the last decade has pioneered the epistemology of data and data science, and particularly the “relational” approach to data, which recognises that data do not speak for themselves and do not drive knowledge creation without considerable human ingenuity and effort.

This approach brings data ethics and human decision-making at the heart of data science and artificial intelligence. For instance, I demonstrated through several detailed case studies in the biological, biomedical and environmental sciences how ethical and secure handling of research data boosts the scientific value of data as evidence, and investment in data infrastructures and community engagement around data results in robust, reliable knowledge. Through studies of Open Data and COVID research, I also demonstrated that data governance – that is, decisions around who has the power to determine what constitutes relevant data and who has access to such data – is crucial to any data-intensive research effort.

Tags: Data Studies, Big Data, Philosophy, Evidence, Good Science, Egenis, Data Ethics.