Exploring Human Mobility: A Time-Informed Approach to Pattern Mining and Sequence Similarity

Area of Science:

• Spatial data science
• Human mobility research
• Computational social science

Background:

• The increasing volume of spatial big data fuels interest in human mobility patterns.
• Discovering and comparing these patterns from big data presents significant analytical challenges.

Purpose of the Study:

• To introduce novel methods for discovering and assessing human mobility patterns.
• To present an analytical framework for analyzing mobility at individual and aggregated levels.
• To demonstrate the framework's application and effectiveness using a real-world case study.

Main Methods:

• Developed Time-Informed pattern mining (TiPam) for frequent pattern discovery.
• Introduced a Time-Aware Longest Common Subsequence (T-LCS) algorithm for sequence similarity assessment.
• Integrated these into a comprehensive analytical framework for human mobility analysis.

Main Results:

• Applied the framework to daily mobile phone data from 135 users in Kampala, Uganda.
• Identified four distinct mobility groups: "stay-at-home," "unoccupied," "education-oriented," and "work-oriented."
• Demonstrated the framework's efficiency and the utility of the novel algorithms.

Conclusions:

• The proposed framework and algorithms effectively analyze human mobility patterns.
• The approach is versatile and applicable to diverse datasets and research fields.
• Provides a robust method for understanding individual and group mobility behaviors.