TelCoVis: Visual Exploration of Co-occurrence in Urban Human Mobility Based on Telco Data

Area of Science:

• Urban studies
• Data visualization
• Human mobility analysis

Background:

• Understanding urban human mobility co-occurrence is crucial for urban planning, business intelligence, social behavior analysis, and disease containment.
• Mobile phone data offers large-scale, fine-grained insights into human mobility but lacks systematic analysis methods.
• Existing methods struggle with in-depth analysis and extracting valuable information from complex co-occurrence patterns.

Purpose of the Study:

• To present TelCoVis, an interactive visual analytics system designed to analyze co-occurrence in urban human mobility using telecommunication (telco) data.
• To enhance analysts' perception and enable comprehensive exploration of human mobility patterns through integrated visualization techniques.
• To facilitate the study of correlations in co-occurrence and identify coordinated relationships among different regions using biclustering.

Main Methods:

• Development of TelCoVis, an interactive visual analytics system integrating novel visualization designs.
• Application of biclustering techniques to analyze correlations in multi-region co-occurrence patterns.
• Case studies using a real-world dataset and validation through interviews with domain experts.

Main Results:

• TelCoVis effectively aids analysts in gaining insights into urban human mobility co-occurrence.
• The system enhances the exploration of coordinated relationships and identification of patterns among different regions.
• Demonstrated effectiveness in facilitating various analytical tasks related to human mobility.

Conclusions:

• TelCoVis provides a powerful tool for analyzing complex urban human mobility data.
• The system supports domain experts in leveraging their knowledge for deeper insights into mobility patterns.
• The integration of visualization and biclustering offers a novel approach to understanding human mobility co-occurrence and correlations.