基于社交媒体签到数据的空间网络及其社区的无标度与热点分析

摘要

In this study, a series of time-dependent spatial interaction networks were firstly derived from massive social media check-in data.Then a geographic weighted community detection method was proposed, and the individual communities were extracted from the spatial interaction networks.Finally, the hot spot and cold spot communities were identified and visualized through the correlation analysis between the community size and activity.The results show that no matter the spatial interaction networks or the community size and activity, they all have scale-free characteristics.It demonstrates the complexity of human activities and their residential environment, implying that the human activities are restricted by the potential geographical environment.The result gives better understanding on the correlation between human beings and the geographical environment of their living cities, which can provide decision supports for urban management and land use planning.%利用大量的社交媒体签到数据构造了不同时间粒度的空间交互网络, 提出了一种地理加权的社区提取算法, 并提取了网络中的社区, 通过社区大小与活跃度的关系可视化地分析并研究了热点与冷点社区.研究表明, 不论是空间交互网络, 还是社区大小与活跃度, 均具有结构的无标度性, 一定程度上说明了人们的活动及其地理环境的复杂性, 暗示了人们的活动受制于潜在的地理环境.研究成果有助于更好地理解人类与城市地理环境的交互关系, 为城市管理和土地利用规划提供决策支持.