Exploration of the spatial pattern of urban residential land use with geographically weighted regression technique: a case study of Nanjing, China

As the traditional methods and technical means cannot meet the quantitative research needs of the urban land use patterns, quantitative research methods for the urban land use pattern are established via the GIS (geographic information system) technique combined with the related theories and models. Taking the city of Nanjing as an example, a spatial database of urban land use and other environmental and socio-economic data is constructed. A multiple linear regression model is developed to determine the statistically significant factors affecting the residential land use distributions. To explain the spatial variations of urban land use patterns, the geographically weighted regression (GWR) is employed to establish spatial associations between these significant factors and the distribution of urban residential land use. The results demonstrate that the GWR can provide an effective approach to the exploration of the urban land use spatial patterns and also provide useful spatial information for planning residential development and other types of urban land use.