Revealing spatiotemporal interaction patterns behind complex cities

Area of Science:

• Complex Systems Science
• Urban Dynamics
• Computational Social Science

Background:

• Cities are dynamic complex systems with intricate human interaction networks.
• Understanding collective spatiotemporal interaction patterns is vital for urban studies.
• Existing research lacks a comprehensive understanding of these urban dynamics.

Purpose of the Study:

• To reveal general collective patterns in spatiotemporal interactions of city residents.
• To analyze urban dynamics using massive cellphone data and network analysis.
• To develop a predictive model for human mobility and urban interaction patterns.

Main Methods:

• Construction of interaction networks using spatiotemporal co-occurrence from cellphone data.
• Analysis of rank-size distributions of dynamic urban populations.
• Aggregation of spatiotemporal interaction networks to identify city state switching.
• Development and application of a temporal-population-weighted-opportunity model.

Main Results:

• Stable rank-size distributions observed in dynamic urban populations across time windows.
• Cities exhibit switching behavior between 'active' (concentrated) and 'inactive' (scattered) states.
• Larger cities show stronger heterogeneity, indicated by a higher scaling exponent.
• A city's active state duration correlates positively with its population size.
• Spatiotemporal interaction patterns are approximated by residential patterns only in smaller cities.

Conclusions:

• The study reveals universal patterns of urban dynamics and human interaction across diverse cities.
• The proposed model reasonably explains observed spatiotemporal interaction patterns and human mobility.
• Findings provide crucial insights for urban planning and understanding complex urban systems.