The Impact of Immersion on Cluster Identification Tasks

Area of Science:

• Information Visualization
• Human-Computer Interaction
• Virtual Reality

Background:

• Head-mounted displays (HMDs) enable immersive virtual reality environments (VREs) for data exploration.
• Previous research indicates immersion positively impacts learning and therapeutic outcomes in various fields.
• The applicability of immersion benefits to information visualization (InfoVis) tasks requires investigation.

Purpose of the Study:

• To assess the impact of immersion on cluster identification in scatterplot visualizations.
• To compare user performance across different visualization design spaces, varying in immersion levels.
• To evaluate the effectiveness of virtual reality environments for information visualization tasks.

Main Methods:

• A quantitative user study with 18 participants in a within-subjects design.
• Comparison of four visualization conditions: 2D scatterplot on screen, 3D scatterplot on screen, 3D scatterplot in VRE (miniature), and fully immersive 3D scatterplot in VRE.
• Data collection on accuracy, efficiency, memorability, sense of orientation, and user preference.

Main Results:

• Task performance varied significantly across the four visualization design spaces.
• 2D scatterplots on screens performed worse than 3D visualizations across measured variables.
• Increased immersion in virtual reality environments demonstrated substantial benefits for 3D data and cluster detection.

Conclusions:

• Immersion in virtual reality environments offers significant advantages for information visualization tasks, particularly with 3D data.
• 3D visualizations, especially within immersive VREs, outperform traditional 2D screen-based visualizations for cluster identification.
• Future InfoVis research should explore the potential of VR for complex data analysis and user experience.