Rhythm of the streets: a street classification framework based on street activity patterns

Area of Science:

• Urban Planning
• Mobility Data Analysis
• Geospatial Science

Background:

• Streets are vital for urban development, access, and interaction.
• Current street classifications are often too broad, focusing on function or land use instead of actual activity.
• Understanding dynamic street activities is crucial for effective urban design and management.

Purpose of the Study:

• To propose a novel activity-based street classification framework.
• To categorize street segments based on temporal activity patterns derived from mobility data.
• To offer a more granular and dynamic approach to street classification.

Main Methods:

• Developed an activity-based street classification framework.
• Utilized high-resolution, de-identified mobility data.
• Applied the framework to 18,023 street segments in Boston.

Main Results:

• Identified 10 distinct activity-based street types (ASTs).
• Demonstrated that ASTs capture dynamic street activities, complementing static classifications.
• Showcased improved identification of streets with higher crime prevalence compared to traditional methods.

Conclusions:

• Activity-based street classification provides finer granularity than existing methods.
• This framework offers valuable insights for urban management and planning.
• Dynamic street activity patterns are key to understanding and improving urban environments.