Understanding analytical results in visual data analysis is essential to inform further exploration and decision-making. However, existing tools often visualize the results alone, offering limited access to associated contexts needed to explain why the results occur. To address this gap, we propose CAMV, a framework that automatically extracts relevant contextual information from given results and datasets, and generates multi-view visualizations to support deeper understanding. CAMV consists of two major components: one that identifies key data subspaces and fact types as explanatory context, and the other that generates coordinated multi-view visualizations of both the analysis results and their contextual information. Based on the framework, we developed an interactive prototype and conducted a comparison study with 12 participants, using GPT-4o as a baseline. Results show that CAMV enhances understanding, decision-making speed, and trust compared to GPT-4o, with positive feedback on its data-grounded reasoning and structured output.
Lin et al. (Wed,) studied this question.