As machine learning systems are rapidly changing our world, the role of data science as a discipline is elevated. However, existing approaches in science communication and data visualization are not sufficient to address the complexity of such systems and the profoundly changing relationship between humankind and data. We introduce the “Data Art & Science” (DAS) project in which the role of art as an integral part in the discourse about data is inves- tigated through a series of commissioned art projects and close monitoring of their creation process and impact. For this paper, we propose a Data Art Research model and use it to discuss one of these art projects, 23nm. We aim to show the value in creating a system to reason about DAS and position the latter as a new way of approaching data-reliant projects in academia and industry to increase the quality of public knowledge and discourse.