Spatial Non-Stationarity And Heterogeneity Of Metropolitan Housing Prices: The Case Of Guangzhou, China

Taking a case study of the city of Guangzhou, this paper applies sophisticated approaches, including spatial autocorrelation analysis models, traditional regression models and Geographically Weighted Regressing (GWR) model to analyse and discuss the spatial non-stationarity and heterogeneity characteristics of housing prices, in term of the data of the average annual housing prices of each sub-district/town from 2009 to 2015. In this study, four types of influence factors, i.e., location condition (distance to central business district (CBD)), traffic access level, commercial service facility and public service facility are selected. As results, the spatial-temporal evolution characteristic of the housing prices in Guangzhou is closely related to the layout and development of the urban areas. Furthermore, among the selected factors, location condition is the most important factors affecting the housing prices, followed by commercial service facility, traffic access level and public service facility. Specifically, based on the GWR model analysis, the spatial patterns of the regression coefficients of location condition and the public service facility illustrate "east strong and west weak" of the housing prices in Guangzhou. By contraries, "west strong and east weak" of housing prices according to the traffic access level and commercial service facility. The overall pattern of spatial non-stationarity and heterogeneity of housing prices presents the circle structure characteristics with the centre of the CBD in Guangzhou.