Interactive History Support for the Exploratory Design of Data Visualizations

Abstract

Research into the design process when building data visualization has yielded many tools that support linear versioning techniques, such as annotation, bookmarking, and interactive history tracking. However, visualization design processes are usually non-linear, and support for exploratory techniques such as branching to build on past iterations and merging operations are underexplored. In this thesis, I investigate how to adapt these related techniques to support the exploratory design phase when designing and building data visualizations. I also introduce novel approaches for merging on a subcomponent level, that allow designers to merge either axes, marks, or legend properties individually. This is motivated by the need to cherry pick different aspects or components of visualizations created at different points of the exploration process to create new ones. At the end, I present a preliminary implementation to support and enhance the exploratory design of data visualizations using Lyra 2, a visual design environment. I also present an evaluation, in the form of supported use cases, which assesses the impact of these features on the exploratory design phase when designing and building data visualisations.