Discovering spatial co‐location patterns of urban facilities and their asymmetric characteristics

Abstract

Spatial co‐location pattern (CP) mining can discover sets of geographical features frequently appearing in adjacent locations, which is valuable for comprehending the co‐occurrence relationship between features. However, due to the quantitative differences and heterogeneous distribution of features, the probabilities that features appear in each other's neighborhood are unequal, resulting in an asymmetric spatial pattern. Current studies have paid little attention to the asymmetric characteristics of CPs. Therefore, this study explores the CPs and their asymmetric relationships. Firstly, we adopt the weighted participation index to evaluate the frequency of global candidate CPs. Secondly, we employ an asymmetry index we developed and the local co‐location quotient to quantify the asymmetry intensity of CPs. The results indicate that the frequent CPs mainly comprise facilities related to the residents' daily lives. Investigating the asymmetric relationships and spatial associations among features in the CPs is significant for identifying resource shortages and rationally planning urban resources.