SenseMap: Urban Performance Visualization and Analytics Via Semantic Textual Similarity.

Area of Science:

• Urban Analytics
• Geospatial Information Science
• Computational Social Science

Background:

• Growing urban populations necessitate better measures of urban performance like livability and comfort.
• Point of Interest (POI) data is widely used in location-based services but underexplored for urban performance analytics.

Purpose of the Study:

• To introduce SenseMap, a novel approach for urban performance analysis using POI data.
• To leverage POI data as a semantic representation of urban functions for performance measurement.
• To develop a visual analytics system for exploring urban performance.

Main Methods:

• Quantifying POI contributions to urban performance measures via semantic textual similarity.
• Proposing Semantic-adaptive Kernel Density Estimation considering POI influence and semantic contributions.
• Generating semantic density maps for various urban performance measures.

Main Results:

• Evaluations confirm the feasibility and validity of the SenseMap method using human judgment and reference data.
• The system demonstrates capability, usability, and explainability through usage scenarios and user studies.
• Semantic density maps provide a novel way to visualize and understand urban performance.

Conclusions:

• SenseMap offers a viable and effective method for analyzing urban performance using POI data.
• The semantic representation of urban functions enhances traditional urban analytics.
• The developed visual analytics system empowers users to explore and understand urban performance in their surroundings.