The Morphology and Circuity of Walkable and Drivable Street Networks

TL;DR

This study compares walkable and drivable street network circuity in 40 US cities, revealing that walking networks generally provide more direct routes than driving networks, highlighting the importance of mode-specific network analysis.

Contribution

It introduces a comparative analysis of walkable versus drivable network circuity using OpenStreetMap data and simulates four million routes across multiple cities.

Findings

Walking networks are more direct than driving networks in most cities.

Average driving circuity exceeds walking circuity in all but four cities.

Network circuity varies significantly between different urban areas.

Abstract

Circuity, the ratio of network distances to straight-line distances, is an important measure of urban street network structure and transportation efficiency. Circuity results from a circulation network's configuration, planning, and underlying terrain. In turn, it impacts how humans use urban space for settlement and travel. Although past research has examined overall street network circuity, researchers have not studied the relative circuity of walkable versus drivable circulation networks. This study uses OpenStreetMap data to explore relative network circuity. We download walkable and drivable networks for 40 US cities using the OSMnx software, which we then use to simulate four million routes and analyze circuity to characterize network structure. We find that walking networks tend to allow for more direct routes than driving networks do in most cities: average driving circuity exceeds average walking circuity in all but four of the cities that exhibit statistically significant differences between network types. We discuss various reasons for this phenomenon, illustrated with case studies. Network circuity also varies substantially between different types of places. These findings underscore the value of using network-based distances and times rather than straight-line when studying urban travel and access. They also suggest the importance of differentiating between walkable and drivable circulation networks when modeling and characterizing urban street networks: although different modes' networks overlap in any given city, their relative structure and performance vary in most cities.