Associations of Street Network and Urban Form with Walking Frequencies

Abstract

Walking is globally promoted as an urban transport mode that reduces greenhouse gas emissions and increases physical activity. While it is well-known that accessibility at neighbourhood and urban levels impact walking levels, studies on pedestrian travel behaviour are typically based on area-based measures. Network-based measures of street configuration, connectivity, urban density, and land-use diversity have scarcely been applied. This paper aims to find associations of network-based connectivity and accessibility measures with walking frequencies. We applied a series of multivariable ordinal regressions adjusted for distance to various destinations, building-type diversity, socio-demographic variables, and car ownership. We operationalise connectivity as reach, i.e., the number of streets reached through the network, and density and diversity as weighted reach, i.e., the floor area, population, or retail accessible within walkable distances through the network. Street network configurations are described by topological centrality (integration). Results show significant positive associations of reach, attraction reach, and integration with walking frequency, with the best model fits achieved by accessible total floor area and reach within relatively high network radii (1000 – 2000 m and 6 axial steps), indicating that connectivity and access to a variety of attraction and functions, often beyond administrative borders, encourage walking.