Comparison of the distributions of centrality indices: Using spatial big data to understand urban spatial structure.

This article seeks to estimate various centrality indices based on spatial big data in terms of urban form, function, and location potential within a metropolitan area; it also comparatively analyzes the morphological differences in urban spatial structure between the centrality indices. As a result of estimating the centrality index for each dimension related to the urban spatial structure of the Seoul metropolitan region based on big data, the differentiated distribution patterns of urban centers were confirmed according to the characteristics of available big data by a quantitative and three-dimension morphological comparison methodology. The Mantel tests showed statistically significant mutual correlations between all centrality indices; however, the Passing–Bablok regressions revealed that the location potential centrality defined as Hansen-type accessibility was significantly different from the two centrality indices of the as urban form and function dimensions. The centrality of various dimensions in this paper includes not only the spatial characteristics of detailed spatial units but also the temporal characteristics theoretically. These results imply the need for comprehensive research to be carried out on urban spatial structure and centrality in accordance with the polycentric pattern of spatial structure in metropolitan areas. • Centralities in various dimensions were estimated based on spatial big data. • Centrality estimation methods are spatial convolution and accessibility calculation. • Distribution of city centers was different by urban form, function, and potential. • Pattern of location potential centrality was differentiated from other centralities. [ABSTRACT FROM AUTHOR]