A Theoretical Approach for Structuring and Analysing Knowledge Provenance for Visual Analytics

The primary goal of Visual Analytics (VA) is to enable user-guided knowledge generation. Theoretical VA works to explain how the different aspects of a VA tool bring forth new insights through user interactivity, which itself can be captured through tracking methods for reproduction or evaluation. However, the process of automatically capturing the user's thought process, such as intent and insights, and associating it with user's interaction events are largely ignored. Also, two forms of interactivity capture are typically ambiguous and intermixed: the temporal aspect, which indicates sequences of events, and the atemporal aspect, which explains the workflow as sequences of states within a state-space. In this work, we propose Visual Analytics Knowledge Graph (VAKG), a conceptual framework that brings VA modeling theory to practice through a novel Set-Theory formalization of knowledge modeling. By extracting such a model from a VA tool, VAKG structures a 4-way temporal knowledge graph that describes user behavior and its associated knowledge gain process. Such knowledge graphs can be populated manually or automatically during user analysis sessions, which can then be analyzed using graph analysis methods. VAKG is demonstrated by modeling and collecting Tableau and visual text-mining workflows, where comparative user satisfaction, tool efficacy, and overall workflow shortcomings can be extracted from the knowledge graph.