Towards Understanding the Impact of Guidance in Data Visualization Systems for Domain Experts

TL;DR

This paper investigates how guidance affects interpretation and communication when domain experts use data visualizations. The authors develop visAPPprot, a coarse-to-fine guided visualization system for proteomics, and conduct two user studies with domain experts using a fixed expression dataset to compare guided versus unguided workflows. Key findings show that guided, coarse-to-fine analysis improves both interpretation of visualizations and storytelling of insights, guiding users from high-level overviews to detailed relationships. The work provides empirical evidence and design considerations for integrating guidance into visualization systems to better support domain-specific exploration and communication in complex biomedical data.

Abstract

Guided data visualization systems are highly useful for domain experts to highlight important trends in their large-scale and complex datasets. However, more work is needed to understand the impact of guidance on interpreting data visualizations as well as on the resulting use of visualizations when communicating insights. We conducted two user studies with domain experts and found that experts benefit from a guided coarse-to-fine structure when using data visualization systems, as this is the same structure in which they communicate findings.

Figures (3)

• Figure 1: Panels a), b), and c) together demonstrate use of visAPPprot visualizations to communicate findings about input dataset.
• Figure 2: Top: Behaviors exhibited by participants. Bottom: Most and least intuitive visualizations.
• Figure 3: Visualization interpretation by guided and unguided cohort, separated by category and plotted by individual participant results.