The widespread adoption of machine learning and other automated forms of reasoning in scientific practice demands careful philosophical reflection. I am interested in tackling broad philosophical questions concerning the epistemology of these techniques within data-driven science. What are the aims of data-driven science, and how can these distinct aims help us to evaluate novel methods? Are machine learning systems models, or something else? How, if at all, do machine learning systems explain? Can they generate improved understanding of the objects of scientific investigation? My answers to these questions lay the groundwork for more specific investigations into the moral and epistemological significance of particular data-driven techniques, methods, and activities within scientific practice.