Evaluating Perceptually Complementary Views for Network Exploration Tasks
summary
Three controlled studies evaluate matrix, node-link and MatrixWave versus Sankey on weighted and event‑sequence networks with 29 and 24 participants respectively, plus eye tracking. Key findings: an eight‑way pretest picks square encoding for matrix cells; a combined side-by-side view with brushing and linking matches the accuracy of the best single view but increases completion time; 56% of users stick exclusively to one view after training while ~11% switch rapidly, showing strong view preference.
AI-written summary · about
chapters (8)
- 0:00Introduction to research
- 2:37Motivation for combined views
- 4:26Perceptual complementarity definition
- 6:43Pre-study on matrix encodings
- 9:20Study 1: Node-link vs matrix
- 12:56Study 2: Matrix wave vs Sankey
- 15:31Summary of findings
- 16:58Q&A session
related talks
02h 47mNetwork Analysis Made Simple
43mVisualising networks
19mAn Eye For Design: Gaze Visualizations for Remote Collaborative Work
11mA Comparison of Cooperative and Competitive Visualizations for Co-located Collaboration
06mMultivariate Visualization of Game Metrics: An Evaluation of Hexbin Maps
42mInteractive Data Analysis: Visualization and Beyond
46mExplorations in Interactive Visual Analytics
42mVisualising Connections