Understanding the spatial organization of urban functions based on co-location patterns mining: A comparative analysis for 25 Chinese cities

A proper understanding of urban functions is fundamental to prevent urban problems and promote better built environments. While previous studies focus mainly on inferring urban function types, there are few understandings of how urban functions organize spatially from the perspective of spatial co-occurrence of activities. To address this issue, we propose a new method to mine the co-location patterns (CPs) of urban activities with ‘Point-of-Interest’ (POI) data. We implement a comparative analysis with 25 major cities in China to recognize their commonness/distinctiveness in the spatial organization of urban functions. We identify nearly a thousand unique POI CPs for these cities. By aggregating the resulting CPs, we create the urban function graphs for each city to reflect the inter-connections of different function types. Most cities have relatively high graph densities, suggesting the mixture of urban functions in the space. Despite the relatively similar structures among cities, their manifestations of POI CPs vary greatly from one city to another. The commonly found POI CPs (witnessed in at least 20 cities) only contributes to 6.2% of the total unique POI CPs. The cities of Wuxi, Chengdu, Ningbo, Shenzhen and Chongqing have most distinctive urban functional structures compared to other cities.