Data structure choices impose constraints on the kinds of questions that we can ask, at least in terms of query and runtime complexity. In a visualization design study, we normally determine the data format and the data abstraction before choosing visual encodings—however, in practice, we often have to rework and transform the data as soon as a user asks a new question. Additionally, outside of design studies, analysts must reshape data on their own: few visualization systems have the flexibility to reshape data on the fly, making it necessary to wrangle the data before loading it into a visualization tool.
Existing tools like Excel support reshaping tabular data, however, richer reshaping operations are possible in programming environments. Our goal in this project is to make some of these operations—specifically graph reshaping operations—more accessible to practitioners, without the need to learn and write code. In this week's visualization seminar, I will be soliciting feedback on our overall study design, suggestions for reshaping tasks that we may have overlooked, and provide a preview of the system that I will be implementing.