UnDRground Tubes: Exploring Spatial Data with Multidimensional Projections and Set Visualization

Area of Science:

• Data Visualization
• Geostatistics
• Scientific Computing

Background:

• Multivariate spatial data analysis is crucial in scientific and industrial fields.
• Spatial Blind Source Separation (SBSS) is a powerful technique for analyzing such data, outperforming non-spatial methods like PCA.
• The complexity of latent components in SBSS hinders effective analysis, especially with varying parameter settings.

Purpose of the Study:

• To address the challenge of analyzing complex latent components in SBSS.
• To propose a novel visualization approach, UnDRground Tubes (UT), for enhanced spatial data analysis.
• To integrate UT into an interactive system and evaluate its effectiveness.

Main Methods:

• Developed UnDRground Tubes (UT), a visualization idiom combining set visualization and multidimensional projections.
• Integrated UT into an interactive multiple-view system.
• Conducted interviews with SBSS experts, qualitative evaluations with visualization experts, and computational experiments.

Main Results:

• SBSS experts expressed enthusiasm for UT, recognizing its benefits for their work and broader geostatistical applications.
• Visualization experts positively received the UT approach.
• Computational benchmarks confirmed the appropriateness of UT projections and heuristics.

Conclusions:

• The proposed UnDRground Tubes (UT) visualization approach effectively tackles the complexity of latent components in SBSS.
• UT offers significant advantages for multivariate spatial data analysis and geostatistics.
• The interactive system integrating UT is well-received and validated by experts.