Characteristics and Driving Factors of Industrial Activity Spatial Distribution in Mega-Cities from the 'Land-Building' Perspective: A Case Study of Shenzhen

Megacities serve as primary spatial carriers of industrialization and sources of innovation and upgrading. Existing research predominantly relies on business registration data to identify the micro-locations of industrial activities; however, this approach suffers from limitations such as low identification accuracy and difficulties in subdividing the types of industrial activities. This study utilizes Shenzhen as a case and constructs a new method for identifying the locations of industrial activities based on the "Land-Building" perspective. By integrating land use and building data and by employing kernel density estimation, spatial autocorrelation analysis, and spatial econometric models, this study systematically examined the spatial distribution characteristics and factors influencing industrial production and R&D activities. The findings reveal that 1) the overall spatial distribution of industrial activities in Shenzhen exhibits a pattern of "dense in the north and sparse in the south," with 95.06% of industrial activity buildings located in the six districts outside the "Guan" area and 4.94% in the four districts inside the "Guan" area. There are significant differences between the spatial locations of industrial and industrial production activities. Industrial production activities display a "multicenter contiguous" layout, primarily distributed in the northeastern "Longgang-Pingshan" industrial production zone, the northwestern "Bao'an-Guangming" production zone, and the cross-city production zone at the border between Shenzhen and Dongguan. In contrast, industrial R&D activities exhibit a "single-center scattered" spatial distribution, mainly concentrated in central urban areas such as Nanshan District and Futian District, with key clusters in streets such as Yuehai, Xili, and Xixiang. 2) Both urban industrial R&D and production activities exhibited positive spatial autocorrelation and spatial spillover effects. Moran's I index for urban industrial R&D is 0.518, while that for production activities is 0.123, indicating that the spatial autocorrelation and spillover effects of industrial production activities are stronger than those of industrial R&D activities. Industrial production activities can exert economic agglomeration effects over a broad geographical range, whereas industrial R&D activities generate agglomeration effects only in smaller areas. The spatial clustering of industrial production activities centers around Fubao Street, with low-low clustering dominating the fan-shaped area from Xixiang Street to Yantian Street and high-high clustering prevailing outside this area. The spatial clustering of industrial R&D activities exhibits an "interlaced and embedded" pattern, forming a nested spatial distribution. 3) The spatial layout of industrial activities in megacities is the result of the combined effects of multiple factors including socioeconomic development, transportation infrastructure, production and living facilities, and geographical location. However, the driving factors for the microlocation distribution of the two types of activities differ. Factors such as population density, industrial development, highways, railway stations, commercial facilities and administrative area size positively influenced the spatial agglomeration of industrial production activities. In contrast, the economic level, industrial development, bus stops, subway stations, and financial facilities positively affect the spatial agglomeration of industrial R&D activities. This study suggests that governments should consider the regularity and heterogeneity of the spatial agglomeration of different industrial activities when conducting urban spatial and industrial planning, promote reasonable functional zoning of urban industrial activities, and enhance the supply of industrial development factors to foster the agglomeration and development of industrial activities.