The project plans to analyze available digital resources for the three millennia of available cuneiform written evidence (c. 3,000 BCE - 100 CE). It will use both measurable statistical methods, as well as cutting-edge explainable machine learning models such as variational auto encoders (VAEs), to bridge the gap between macro and microhistory. The goal of the research is to examine case studies situated between the historian's desire to find general explainable patterns and the desire for high resolution evidence of events. The first case study of the project is the analysis of cuneiform tablet shapes and their relationship to historical date, genre, and archival context.