Spatial distribution of access diversity on urban road networks

Abstract

The road network in an urban area plays an essential role in mediating the mobility of people and goods and in supporting various social and economic activities. In urban road networks, accessibility is one of the crucial aspects of evaluating their performance. The existing measures from urban science focus on practical user-based accessibility utilizing various data, including human mobility patterns, transport networks, and land use layout. While such measures can present a comprehensive picture of accessibility, they hardly describe how a single factor, essentially the structural and topological patterns of road networks, determines accessibility level. Meanwhile, the random-walk-based accessibility used for various complex networks can effectively analyze structurewise accessibility. However, we claim that the frequently used version of the random-walk-based accessibility only considers the topology of networks, and thus often neglects important geometric information such as the spatial physical distances in road networks. In this study, we modify the existing random-walk-based accessibility method to be more applicable for road networks by including information of geometric distance. To evaluate the effect of the distance information on accessibility, we analyze the empirical road networks of three global metropolitan cities with different structures. We find that our spatial accessibility captures the disparity in spatial and topological structures and highlights spatially accessible areas in different cities. Especially, our methodology brings a dramatic change in the cities with heterogeneous road patterns. We also compare the spatial distribution patterns of our proposed accessibility in the three cities and discuss how road structures are related to the accessibility distributions.