Doctor of Philosophy

dissertationIteration is a ubiquitous need across many different visualization design workflows. For visual design, practitioners often need to transition between the automated generation of visuals and manual drawing; however, existing software makes free-form iteration between these modalities difficult or impossible. A similar challenge exists for data as well: data abstractions often need to be adjusted; however, many data reshaping operations- particularly graph wrangling operations-have support only in programming environ- ments that may not be accessible to many visualization designers. This dissertation begins with an effort to better understand how graphic designers, a specific subset of the broader visualization community, work with data. The lessons from that effort inspire our remaining contributions, which are directed toward the broader community of visualization practitioners: first, we present a software model and a system that make it possible to iterate between drawing and generative visualization. Second, we present a visual technique that enables inquiry into the relationship between two specific graph data wrangling operations: pivoting and filtering. Finally, we present an interface that supports a broader range of graph data wrangling operations that enable iteration between many different graph data abstractions