Hegel famously called history a slaughter-bench (Schlachtbank); Franco Moretti transferred this metaphor to literature, referring to the “great unread”, an archive of forgotten books to be rediscovered through distant reading. I will borrow it in order to allude to the extensive array of neglected and overlooked sources in the history of science, which could be brought back into view through datafication. These are what I call leftovers of science: drafts, protocols of failed experiments, obsolete scientific instruments, photographs, reports and other outdated documentation of the laboratory routine. Such artifacts constitute by-products of knowledge production, which are left behind by science. They are, however, of crucial importance for the history of science, as they constitute object-knowledge, bearing testimony to certain facets of science’s history, participating in our understanding of “science in the making”. The question is how to (again) make them part of the knowledge process? How to turn them into data readable and interpretable for the history of science? The poster presents some preliminary reflections toward a new data model for the leftovers of science.