Color-Encoded Links Improve Homophily Perception in Node-Link Diagrams

Area of Science:

• Information Visualization
• Network Analysis
• Human-Computer Interaction

Background:

• Node-link diagrams are crucial for understanding network structures.
• Assessing homophily, the principle that similar entities tend to connect, relies on distinguishing intra-cluster and inter-cluster links.
• Existing visualization designs may not optimally support accurate homophily assessment.

Purpose of the Study:

• To investigate the impact of edge color encoding on the visual assessment of homophily in node-link diagrams.
• To compare the effectiveness of a novel design (edge color for link type) against a traditional design (same-color edges).

Main Methods:

• An online experiment was conducted.
• Participants were presented with node-link diagrams featuring different edge coloring strategies.
• Data was collected on the accuracy of homophily perception.

Main Results:

• The new design, encoding link type with edge color, resulted in more accurate perception of homophily.
• This contrasts with designs where all edges share the same color, hindering distinction between link types.
• Viewer ability to identify and evaluate link counts is key to homophily assessment.

Conclusions:

• Encoding link type using edge color significantly enhances the accuracy of visual homophily assessment in node-link diagrams.
• This design improvement aids viewers in distinguishing between intra-cluster and inter-cluster links.
• The findings suggest practical implications for the design of network visualization tools.