From Data to Knowledge Graphs: A Multi-Layered Method to Model User's Visual Analytics Workflow for Analytical Purposes

The primary goal of Visual Analytics (VA) is knowledge generation. In this process, VA knowledge models and ontologies have shown to be beneficial to better understand how users obtain new insights when executing a VA workflow. Yet, the gap between theoretical models and the practice of knowledge generation analysis is wide, and theory has mainly been used as a baseline for practical works. Also, two concepts are typically ambiguous and intermixed when analyzing VA workflows: the temporal aspect, which indicates sequences of events, and the atemporal aspect, which indicates the workflow's state-space, which is the set of all states of the VA tool and its user occupied during a VA workflow. Also, the lack of guidelines on how to analyze the recorded user's knowledge-gathering process when compared to the VA workflow itself is apparent. We bridge this gap by presenting Visual Analytics Knowledge Graph (VAKG), a conceptual framework to bridge the gap between VA workflow modeling theory and application. Through a novel Set-Theory formalization of knowledge modeling, VAKG structures a VA workflow by temporal sequences of human and machine changes over time and how they relate to the workflow's state-space. This structure is then used as a schema for storing VA workflow data and can be used to analyze user behavior and knowledge generation. VAKG is designed following the needs and limitations of relevant literature, allowing for modeling, structuring, storing, and providing analysis guidelines for user behavior and knowledge generation, enabling comparison of users and VA tools.