Investigating Visual Perception of Degree Centrality in Graph Visualization.

Area of Science:

• Graph theory
• Human-computer interaction
• Data visualization

Background:

• Degree centrality (DC) quantifies node importance in data space.
• Node-link diagrams visualize graphs, aiding perception in visual space.
• Limited research connects computed importance with visual perception in graphs.

Purpose of the Study:

• Investigate the relationship between computed and perceived node importance.
• Assess accuracy of visual estimation of relative degree centrality.
• Identify visual factors influencing degree centrality perception.

Main Methods:

• A controlled user experiment was designed.
• Participants estimated relative degree centrality from node-link diagrams.
• Analysis focused on accuracy and influencing visual factors.

Main Results:

• Participants inaccurately estimated relative degree centrality, especially with small differences.
• Seven visual factors identified: node receptive region size, local link/node density, neighbor wrapping angle.
• Factor influence varied in complex graph scenarios.

Conclusions:

• Visual perception of degree centrality is often inaccurate.
• Understanding visual factors can optimize graph visualizations.
• Findings bridge computational graph analytics and visual perception.