Measuring Urban Metrics of Creativity Using a Grid-Based Geographically Weighted Regression Model

Jinliao He,Xianjin Huang,Guangliang Xi
DOI: https://doi.org/10.1061/(asce)up.1943-5444.0000450
IF: 2.361
2018-01-01
Journal of Urban Planning and Development
Abstract:Recent urban studies have recognized cities as epicenters of a rapid rise in the creative industries and the creative class. Nevertheless, the relationship between urban spaces and creative industries has yet to be fully understood. This is especially true in places in the emerging world such as China, where the social context is different from that in developed countries. This paper attempts to measure urban metrics of creativity in the city of Nanjing, China, by establishing grid-based spatial regression models. A variety of location factors regarded as crucial for the development of creative industries by previous studies are investigated, including heritage sites, landscape amenities, universities, nightlife, and catering places, as well as public transport. The results suggest that these urban metrics of creativity bear a significant explanatory power for the location choice of creative enterprises in Nanjing, according to both ordinary least-squares (OLS) and geographically weighted regression (GWR). In fact, results for GWR are significantly stronger than for OLS. Meanwhile, great spatial variations are observed across these variables in the grid-based GWR model. That is, a high significance is found in downtown and highly urbanized areas for the factors of heritage sites, nightlife areas, and universities, whereas the other factors (e.g.,landscape amenities) are relevant only in specific locations in the suburban areas of Nanjing. This study thus proves that spatial regression models are more applicable for measuring urban metrics of creativity than conventional economic models are. This paper also bears some implications for urban policy of creative city strategy. (C) 2018 American Society of Civil Engineers.
What problem does this paper attempt to address?