Reflexivity in Issues of Scale and Representation in a Digital Humanities Project

In this paper, we explore issues that we have encountered in developing a pipeline that combines natural language processing with data analysis and visualization techniques. The characteristics of the corpus - being comprised of diaries of a single person spanning several decades - present both conceptual challenges in terms of issues of representation, and affordances as a source for historical research. We consider these issues in a team context with a particular focus on the generation and interpretation of visualizations.