2D, 2.5D, or 3D? An Exploratory Study on Multilayer Network Visualisations in Virtual Reality.

Area of Science:

• Network Science
• Information Visualization
• Human-Computer Interaction

Background:

• Multilayer networks (MLNs) model complex relational data using interconnected layers.
• The visual arrangement of these layers significantly impacts network readability, an area requiring further investigation.
• Higher-dimensional arrangements (beyond 2D) are common and may benefit from stereoscopic displays.

Purpose of the Study:

• To investigate the impact of different visual arrangements (2D, 2.5D, 3D) on the readability of multilayer networks.
• To evaluate performance across a range of network analysis tasks, from path finding to pattern identification.
• To provide empirical-based recommendations for effective MLN visualization strategies.

Main Methods:

• A human subject study was conducted using a Virtual Reality headset.
• Participants performed six distinct network analysis tasks on MLNs presented in 2D, 2.5D, and 3D arrangements.
• Task performance and user experience were analyzed to assess the effectiveness of each arrangement.

Main Results:

• No single arrangement (2D, 2.5D, or 3D) proved universally superior for all multilayer network analysis tasks.
• Performance varied significantly depending on the specific task being performed.
• The study identified task-specific advantages for different dimensional arrangements.

Conclusions:

• The optimal visualization of multilayer networks is task-dependent.
• Empirical evidence supports tailoring layer arrangements to specific analytical goals.
• Recommendations are provided for the effective use of 2D, 2.5D, and 3D arrangements in MLN analysis.