Study On Spatial Differentiation Of Beijing's Second-hand Housing Prices Based On GWR Model

With the rapid development of urbanization and the continuous evolution of urban spatial structure,the impact of housing price on the national economy and life is becoming more and more important.Space position has become an important factor in housing price,and the spatial analysis method is used in housing price analysis to highlight its unique advantages.At present,there are many studies on housing price and its influencing factors.Most of them use spatial autocorrelation analysis to study the spatial differentiation of price,the Hedonic model study the influence factors,and take the land price and the new housing price as the main research object.The characteristic price model of the general linear regression method ignores the change of the price in the local area.The response is only an average situation in the study area,and it is difficult to reflect the changes in the space,and the results are difficult to be used in the analysis of the urban space structure.Taking the area within Liuhuan road of Beijing as the research object,this paper combines the characteristic price method with the geographically weighted regression(GWR)model to explore the spatial differentiation and the influencing factors of the second-hand house price.First of all,the data collection work is carried out to obtain the actual second-hand housing transaction data.After the data are preprocessed,the exploratory space data analysis method is used to screen the data.Finally,1427 sample datas are selected.After the analysis of the price of second-hand housing in Beijing,spatial interpolation analysis and spatia l clustering analysis,it is found that the second-hand housing price in Beijing has a significant positive spatial autocorrelation,and the spatial agglomeration is obvious.There are some local positive spatial autocorrelation in the local 72.9% observat ion points,and 15.9% of the observation points exist local negative spatial autocorrelation,and very few observation points do not have spatial autocorrelation.This shows that the second-hand housing price data has spatial non stationarity.O n this basis,the OLS method and GWR method are used to establish the characteristic price model respectively.Through the comparison and analysis of the model established by two methods,it is found that the model established by the GWR method has a better fitting effect and can explain the spatial differentiation of the second-hand housing price better.Finally,goodness-of-fit of the GWR model,and 7 significant factors such as the elevator,the age of building,the distance from the house to the center of Beijin g city,the key public primary schools,the third grade& class A hospitals,the subway stations and the express ways are visualized.It is found that the urban spatial structure of Beijing is changing from the traditional concentric circle model to the multi-core model,and the ability of the influence factors on the second-hand housing price vary with the spatial position.