SCORE: A Framework for Quantifying Diegesis in Situated Visualization for Augmented Reality

Paper Title

SCORE: A Framework for Quantifying Diegesis in Situated Visualization for Augmented Reality

Publication Info

• Topic area: Augmented Reality-based Situated Visualization (AR-SV) design and evaluation.
• Keywords: Augmented Reality, Situated Visualization, Diegesis, Spatial Proximity, Concreteness, Coherence, Referential Context, Environmental Context, AR Design Framework.

Background and Problem

• Problem / challenge: Existing frameworks for AR-based Situated Visualization (SV) focus primarily on spatial positioning and context but lack a rigorous model to evaluate the quality of integration into the physical environment. This results in conflation of designs that differ in experiential quality.
• Significance: Addressing this gap is critical for designing AR visualizations that seamlessly blend with the physical world, enhancing usability, immersion, and engagement.
• Motivation and related work: Prior work has explored spatial positioning, context-awareness, and interaction paradigms in SV. However, the concept of diegesis—how intrinsic a visualization appears to its environment—has been underexplored. This paper builds on narratology and game design principles to formalize diegesis for AR-SV.

Solution

• Proposed approach: The SCORE framework quantifies diegesis in AR-based SV across five dimensions: Spatial Proximity, Concreteness, Coherence, Referential Context, and Environmental Context.
• Novelty:
  1. A multi-dimensional framework to evaluate diegesis quantitatively and qualitatively.
  2. Empirical validation through a user study (N = 21) demonstrating high inter-rater reliability.
  3. Identification of five distinct categories of diegetic visualization via cluster analysis.
• Procedure and key techniques:
  • Systematic analysis of 50 SV papers and 67 visualization examples.
  • Development of a 6-point ordinal scale for each dimension.
  • Validation through inter-rater reliability testing (α > 0.80 across dimensions).
  • Principal Component Analysis (PCA) to derive weights for the SCORE formula.
  • Cluster analysis to identify recurring design patterns.

Results

• Concrete findings:
  • Five categories of diegetic visualization were identified: Contextual (SCORE = 4.13), Integrated (SCORE = 3.80), Hypernatural (SCORE = 2.87), Superimposed (SCORE = 1.23), and Proximal (SCORE = 0.65).
  • PCA-derived weights for SCORE dimensions: Spatial Proximity (0.09), Concreteness (0.15), Coherence (0.29), Referential Context (0.14), Environmental Context (0.33).
  • High inter-rater reliability (α > 0.80) confirmed the framework’s robustness.
• Advantage over baselines:
  • Differentiates designs that prior frameworks treated as identical by quantifying diegesis on a continuum.
  • Provides a structured vocabulary for analyzing and comparing SV designs.
• Experiments / evaluation:
  • User study with 21 participants experienced in AR/VR systems.
  • Corpus scoring of 67 visualization examples using the validated rubric.
  • Cluster analysis using UMAP and statistical measures to identify design archetypes.
• Limitations and future work:
  • Framework scoped to AR; future work should adapt it for VR and data physicalizations.
  • Does not account for technical implementation factors like rendering fidelity or tracking accuracy.
  • Subjectivity in evaluating qualitative dimensions remains a challenge; further studies could explore variability across user groups.

Summary

The SCORE framework introduces a novel method for quantifying diegesis in AR-based Situated Visualization, addressing gaps in existing models by evaluating five dimensions: Spatial Proximity, Concreteness, Coherence, Referential Context, and Environmental Context. Empirical validation confirmed its reliability, and cluster analysis revealed five distinct categories of diegetic design. This framework enables researchers and designers to systematically analyze and compare SV designs, offering insights into trade-offs between immersion, intuitiveness, and information bandwidth. Future work could extend SCORE to VR environments and explore its relationship with technical factors like rendering fidelity.