Persistent interaction: A conceptualization of user-generated artefacts in Visual Analytics

Abstract

Visual Analytics (VA) is essential for supporting insight generation and knowledge discovery in complex data analysis tasks. However, traditional approaches often overlook the value of user-generated artefacts — such as annotations, parameterizations, selections, spatializations, and other constructs — encapsulating subjective judgments about data and highly contextualized insights. To address this gap, we propose persistent interaction as a paradigm for formalizing how users’ decisions are embedded within artefacts, ensuring their transferability across analytical contexts. In this paper, we introduce a classification of persistent user-generated artefacts, demonstrating their potential through two case studies. We contribute a framework for understanding persistent interaction, insights into artefact generalizability, knowledge transferability, and guidance enhancement, as well as theoretical implications for VA.