Hybrid Graph Visualizations With ChordLink: Algorithms, Experiments, and Applications.

Area of Science:

• Computer Science
• Data Visualization
• Network Analysis

Background:

• Real-world networks exhibit a dual nature: globally sparse yet locally dense.
• This characteristic challenges traditional, homogeneous visualization models.
• Existing hybrid approaches like NodeTrix offer partial solutions.

Purpose of the Study:

• To introduce ChordLink, a hybrid visualization model for complex networks.
• To address the challenge of visualizing both global network structure and local community details simultaneously.
• To enable interactive exploration of network communities.

Main Methods:

• Developed ChordLink, a hybrid visualization integrating node-link diagrams with chord diagrams.
• Chord diagrams represent dense subgraphs within the global node-link layout.
• Implemented a prototype system for ChordLink visualization.

Main Results:

• ChordLink effectively visualizes dense subgraphs (communities) within a sparse global network structure.
• The visualization allows interactive highlighting of community structures.
• The rest of the network layout remains stable during interaction.
• Case studies on real-world networks demonstrate ChordLink's utility.

Conclusions:

• ChordLink offers a powerful solution for visualizing complex networks with sparse and dense regions.
• The model facilitates a deeper understanding of network communities.
• Algorithmic challenges in implementing ChordLink were addressed.