地球信息科学学报 ›› 2017, Vol. 19 ›› Issue (8): 1049-1059.doi: 10.3724/SP.J.1047.2017.01049

• 地理空间分析综合应用 • 上一篇    下一篇

北京市二手房价格时空演变特征

周湘1,2(), 袁文1,*(), 李汉青1,2, 马明清1,2, 袁武3   

  1. 1. 中国科学院地理科学与资源研究所,北京 100101
    2. 中国科学院大学,北京 100049
    3. 北京理工大学计算机学院,北京 100081
  • 收稿日期:2017-03-13 修回日期:2017-05-03 出版日期:2017-08-20 发布日期:2017-08-20
  • 通讯作者: 袁文 E-mail:zhouxiang@lreis.ac.cn;yuanw@lreis.ac.cn
  • 作者简介:

    作者简介:周 湘(1991-),女,湖南邵阳人,硕士生,主要从事房地产时空大数据方面研究。E-mail: zhouxiang@lreis.ac.cn

  • 基金资助:
    资源与环境信息系统国家重点实验室自主创新项目(O88RA20DYA)

Research on the Spatial and Temporal Evolution Characteristics of the Price of Second-hand Housing in Beijing

ZHOU Xiang1,2(), YUAN Wen1,*(), LI Hanqing1,2, MA Mingqing1,2, YUAN Wu3   

  1. 1. Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of science, Beijing 100101, China
    2. University of Chinese Academy of Sciences, Beijing 100049, China
    3. Computer School, Beijing Institute of Technology, Beijing 100081, China.
  • Received:2017-03-13 Revised:2017-05-03 Online:2017-08-20 Published:2017-08-20
  • Contact: YUAN Wen E-mail:zhouxiang@lreis.ac.cn;yuanw@lreis.ac.cn

摘要:

城市住宅价格时空格局及演变特征是衡量城市房地产市场发展均衡性的重要指标。针对海量的互联网实时房产数据,本文构建了一种长时序时空大数据挖掘方法。首先,利用挂牌数据和成交数据,进行了泛在网络地产数据的可用性验证;其次,提出了“混合像元”的多尺度栅格模型,以构建基于栅格系统的房产统计特征描述,形成了多源网络房产数据融合方法;然后分别采用莫兰指数和地理探测器分析房价的空间自相关性和分异性,并基于P-Bshade和邻近栅格时空插值算法解决了稀疏房产数据的融合与插值问题,构建了长时序房地产时空栅格数据库;最后,以北京六环范围内为研究区域,通过栅格区划算法进行了二手房价格时空演变格局的挖掘分析。

关键词: 二手房价格, 泛在网络房地产大数据, 长时序栅格数据库, 时空演变格局

Abstract:

The temporal and spatial pattern of urban housing price and its evolution characteristics are important indicators of measuring the equilibrium of urban real estate market. Based on the large amount of real time data of the Internet, we constructed a spatio-temporal data mining method using long time series. Firstly, a large number of real estate listing information and transaction data existing in different real estate business website had been obtained by web crawler technology. Secondly, the correlation between housing listing price and transaction price was tested using a linear regression model and the usability of ubiquitous online real estate data had been validated. Thirdly, a multi-scale grid model of mixed pixels was proposed, which was based on the description of the statistical characteristics of the real estate, and the problem of multi-source data fusion was solved. Moran’s I and Geo-detector were used to analyze the geographic spatial autocorrelation and non-homogeneity of housing listing price. The spatial raster database of long term real estate was constructed based on the combination of adjacent spatio-temporal interpolation and P-Bshade interpolation. Finally, the inner part of Beijing Six Ring Road was selected as the study area. We analyzed the spatial-temporal evolution characteristics of second-hand housing prices by grid partition algorithm. Overall, we explored the real time dynamic analysis method of real estate. The results showed that: in the first half of 2016, the growth rate of second-hand housing price was larger, and the latter half of the growth was relatively flat. The spatial distribution of second-hand housing price in Beijing was dominated by a single center pattern. At the same time, there are distributions of high island area. Dongcheng and Xicheng district were the core area of high housing prices, and the magnitude of price volatility was not consistent in different direction. The descending velocity of south of central city was the fastest. The diminishing rate of the north and northwest of central city were the slowest. The house price difference was more remarkable in the center of the city than that in the periphery region of the city.

Key words: second-hand housing price, network real estate big data, long sequential raster database, spatio-temporal evolution pattern