国家中心城市共享住宿的时空分布及影响因素——基于DBSCAN算法的分析

共享住宿的兴起是城市旅游服务业供给侧结构性改革的重要推动力,科学认识其时空格局及驱动因素对城市旅游发展意义重大。通过Airbnb网站获取我国九个国家中心城市的共享住宿房源信息,运用标准差椭圆,DBSCAN聚类算法等空间分析方法分析房源的时空演变格局,并借助地理探测器探讨房源空间分布格局的影响因素。结果表明：（1）2015—2017年,国家中心城市共享住宿的空间格局变化较快,整体处于快速发展期,2017—2019年,空间格局趋于稳定,整体处于稳步增长期;（2）国家中心城市共享住宿的发展具有时空依赖性,聚类中心随时间均不断增加,多数城市的共享住宿逐步向城市四周扩展,东中西三个区域之间聚类中心的时空差异明显;（3）从发展方向和高级别聚类中心数量来看,国家中心城市共享住宿的发展方向存在四周型和“单—多”方向主导型以及“0-1”“X-1”“1-X”三种演变模式;（4）商业、交通、人口等指标因素对国家中心城市共享住宿的空间格局产生较大影响,各区域之间及各城市之间影响因素的作用强度和显著性存在明显差异,其中餐饮、休闲娱乐、购物、住宅小区等成为驱动共享住宿空间格局形成的主要影响因素。

Spatio-temporal distribution and influencing factors of sharing accommodation in central cities of China: Analysis based on DBSCAN algorithm

The rise of sharing accommodation is an important driving force for the structural reform of the supply side of the urban tourism industry. It is of great significance to scientifically understand its spatio-temporal pattern and driving factors for the development of urban tourism. Based on the information on sharing accommodation listings in nine national-level central cities of China through the Airbnb website, this study uses the spatial analysis methods such as standard deviational ellipse and DBSCAN clustering algorithm to analyze the spatial and temporal evolution pattern of houses, and to explore influencing factors of the spatial distribution of housing listings with the geodetector. The results show that: (1) From 2015 to 2017, the spatial pattern of sharing accommodation in national central cities changed rapidly, and the overall situation was in a period of rapid development. From 2017 to 2019, the spatial pattern tended to be stable, which was in a steady growth period. (2) The development of sharing accommodation in national central cities has a spatial and temporal dependence, and the clustering centers continue to increase with time. The sharing accommodation in most cities gradually expands around the city. The spatial and temporal differences between clustering centers in the three regions of the Eastern, Central and Western China are obvious. (3) From the perspective of development direction and the number of high-level clustering centers, the sharing accommodation in national central cities includes four-sided and "single-to-multiple" directions, as well as "0-1", "X-1" and "1-X" evolutionary models. (4) Business, transportation, population and other factors have a greater impact on the spatial pattern of sharing accommodation in national central cities, and there are significant differences in the intensity and significance of the influencing factors between regions and cities. Among them, catering, leisure, shopping, residential quarters, etc. have become the main factors driving the formation of sharing accommodation space.