Spatial inequality of bus transit dependence on urban streets and its relationships with socioeconomic intensities: A tale of two megacities in China

The dependence of urban bus transit on their covered streets is expected to be significant and heterogeneous in megacities. Using a bipartite network approach, we develop several weighted centrality-based connectivities to quantify the degree of dependencies of urban bus transit on streets. Two megacities with different road patterns, i.e., Beijing and Shanghai in China are taken as examples to depict comparatively spatial inequalities of the centrality-based dependencies at both local and global scales. Then, a series of spatial cross-section regression models are introduced to explore the colocation relationships between the dependencies of bus transit and urban socioeconomic intensities. The methodology of kernel density estimation (KDE) is used to convert all data with different scales into the same unit of measurement. Results indicate that there are significant statistical and spatial inequalities of the centrality-based dependencies of bus transit on urban streets. These inequalities with evident hierarchies, variances and clusters were validated by statistical analysis including power-law function, rank-size distribution, multiple variance indices, together with Global Moran's I. Besides, a majority of bus transit rely heavily on minor streets concentrating on circumferential expressways in central urban areas and radial highways oriented to outer suburbs under industrial or residential suburbanization. The unequal distribution is found to be strongly related to population, nighttime light intensity, transport-related services, and commercial and leisure services by the spatial regression models. A good spatial matching between bus routes' dependencies and socioeconomic activities intensities is found both in these two megacities.