A Comparison of Spatiotemporal Visualizations for 3D Urban Analytics.

Area of Science:

• Urban Informatics
• Geographic Information Systems (GIS)
• Data Visualization

Background:

• Increasing availability of 3D urban datasets (e.g., shadow, noise, solar potential, earthquake simulations).
• Need for effective visualizations to analyze dynamic urban behaviors in 3D spatiotemporal dimensions.
• Challenges in designing visual strategies for time-varying data on city geometry.

Purpose of the Study:

• Investigate the effectiveness of different visual designs for 3D spatiotemporal urban data analysis.
• Compare four representative visualization techniques: spatial juxtaposition, temporal juxtaposition, linked view, and embedded view.
• Provide design recommendations for 3D spatiotemporal urban visualizations.

Main Methods:

• Elicited analytical tasks from urban domain practitioners.
• Conducted a quantitative user study comparing four visualization methods.
• Assessed user performance (accuracy and speed) on tasks involving extreme value identification on building surfaces over time.

Main Results:

• Plot-based visualizations (linked, embedded) led to higher accuracy.
• Color-coded visualizations (spatial, temporal juxtaposition) were faster for simpler tasks.
• Plot-based visualizations showed better efficiency preservation with increasing task complexity.

Conclusions:

• The choice of visualization impacts accuracy and speed in 3D spatiotemporal urban data analysis.
• Plot-based designs are superior for complex analyses, while color-coded designs suit simpler tasks.
• Design recommendations are provided for practitioners and researchers working with 3D spatiotemporal urban data.