On the surface, Big Data seems an obvious and rewarding new frontier for historians, with digitisation offering new insights on a grand scale. However, as an observational rather than experimental discipline, and with datasets particularly prone to unfillable gaps owing to labyrinthine copyright claims and an uneven electronic public domain, data-driven historical research can appear either impossible or thoroughly undesirable. Yet, the sentiments behind open data and methodological reproducibility are part of the very fabric of the humanities, with its tradition of documented hermeneutics and thoroughly provenanced evidence. This talk will discuss the historiographical tradition of reproducibility and how we might build upon these rich traditions in an age of data-driven results.