A typology of street patterns

Area of Science:

• Urban morphology and planning
• Computational geography
• Spatial analysis

Background:

• Traditional methods for classifying urban street patterns often lack quantitative rigor.
• Understanding distinct urban layouts is crucial for urban planning and historical analysis.
• A quantitative approach can reveal underlying structures in city development.

Purpose of the Study:

• To develop a quantitative method for classifying cities based on their street patterns.
• To define a 'city fingerprint' using block area and shape factor distributions.
• To create a typology of global cities using hierarchical clustering.

Main Methods:

• Calculating the conditional probability distribution of block shape factors for given areas.
• Defining a city's 'fingerprint' based on these distributions.
• Applying hierarchical clustering to group cities based on their fingerprints.

Main Results:

• Identified four major families of cities based on block area and shape abundance.
• Quantitatively distinguished European and American cities into separate sub-categories.
• Demonstrated that a city's fingerprint aggregates neighborhood-level patterns (e.g., New York boroughs).

Conclusions:

• The proposed quantitative method effectively classifies urban street patterns.
• The 'city fingerprint' provides a novel way to compare urban morphology globally.
• This approach offers insights into the formation and evolution of distinct urban layouts.