Dataset for visitations of public green spaces in Shanghai, China.

GreeMove is a dataset for visitations of public green spaces in Shanghai, China, constructed as a bipartite dynamic mobility network between residential polygons and parks. GreenMove utilizes mobile phone data from a telecom operator in Shanghai from January to April 2014, encompassing 10 million anonymized mobile phone users. We identify individual stays via spatial clustering from raw mobile records and categorize them as home, work, or other activities. We then recognize the other locations those are geographically overlapped to parks to identify the park visitations. And we redefine a division approach for the Shanghai metropolitan area by using cell towers to partition it into Voronoi grids, serving as residential polygons. It can be characterized as a partition with the finest resolution, resulting in a total of 38,055 residential polygons, which are intricately linked to the mobile source data we utilized. GreenMove comprises a daily dynamic network , as well as a more granular daily time segmented network sustained over four months. On a given day or at a given time, if visitation does exist from residential polygons to parks, an edge is established between the two nodes. The daily dynamic network is stored within the "daily_network_exp_8-4_geometry" folder, while the daily time-segmented network is contained in the "daily_segment_network_exp_8-4_geometry" folder. Additionally, a comprehensive super-network spanning a four-month period is saved in the file "4month_network_exp_8-4_geometry.pkl". Furthermore, a CSV file is provided, serving as the input for the GBM model to predict pairwise flow. Code examples for working with the dataset, including the implementation of the GBM model, are available in the GitHub repository.GreeMove是面向中国上海市公共绿地到访行为的数据集，被构建为居住多边形与公园之间的二分图动态移动网络（bipartite dynamic mobility network）。该数据集依托2014年1月至4月上海某电信运营商提供的手机信令数据，涵盖1000万匿名手机用户。研究团队通过对原始手机信令记录进行空间聚类，识别出用户的停留行为，并将其划分为居家、工作或其他活动三类。随后，通过匹配地理重叠区域，识别出与公园重合的其他停留地点，进而确定用户的公园到访行为。 研究团队重新定义了上海大都市区的划分方案：以通信基站为基础，将区域划分为沃罗诺伊网格（Voronoi grids），并将其作为居住多边形。该划分方案具备最高空间分辨率，最终生成38055个居住多边形，且与本研究采用的手机信令源数据紧密关联。 GreenMove包含每日动态网络，以及覆盖四个月的精细化每日分时网络。若在特定日期或时段存在从居住多边形到公园的到访行为，则在两个节点间建立边连接。每日动态网络存储于"daily_network_exp_8-4_geometry"文件夹中，而每日分时网络则存储于"daily_segment_network_exp_8-4_geometry"文件夹内。 此外，覆盖全四个月的综合超网络存储于"4month_network_exp_8-4_geometry.pkl"文件中。此外还提供了一份CSV文件，可作为梯度提升树（Gradient Boosting Machine，GBM）模型的输入数据，用于预测节点对间的流动量。该数据集的使用示例代码（包含GBM模型的实现代码）可在GitHub仓库中获取。