Collective human mobility pattern from taxi trips in urban area

Area of Science:

• Urban planning
• Transportation science
• Data science

Background:

• Understanding urban mobility patterns is crucial for efficient city management.
• Previous studies often lacked the scale to capture nuanced travel behaviors.

Purpose of the Study:

• To identify and quantify the primary purposes of taxi travel in Shanghai.
• To develop a model for approximating urban traffic flow based on identified travel purposes.
• To analyze the variability of traffic patterns and model their deviations.

Main Methods:

• Analysis of 1.58 million taxi trips using non-negative matrix factorization and optimization.
• Development of a linear combination model representing traffic flow using basis flows.
• Statistical modeling of traffic power deviations using probability distribution functions.

Main Results:

• Identified three main workday travel purposes: home-workplace commuting, workplace-to-workplace travel, and other activities.
• Traffic flow can be approximated by a linear combination of three basis flows, termed 'traffic powers'.
• Developed a probability distribution function for traffic power deviations, explained by statistical theories and empirical data.

Conclusions:

• The 'traffic power' model provides a parsimonious way to understand and predict urban traffic flow.
• Findings are applicable to road traffic prediction, pattern tracing, and diagnosing abnormal traffic events.
• The methodology can infer urban land use based on traffic patterns.