Hourly movement numbers and density plot. (a) Hourly movement distribution. The x-axis denotes the 168 hours of the week. The labels in the x-axis follow the pattern of ####-##-##T##+08, following ISO 8601 format (YYYY-MM-DDTHH+08) and inferring the HH hours on YYYY-MM-DD. Y-axis represents the total hourly movement number of the 4 studied places between locations in an hour. (b) Trip duration distribution. The x-axis denotes the 24 hours of trip duration. The y-axis denote the proportion of trip number over all across trip durations. (c) Empirical d egree distribution with two separate fits: (1) A Gamma distribution for large degree values in the Songliao dataset, (2) A Gamma distribution for small degree values in the dataset of Shenzhen taxi passengers. The x-axis denotes the logarithmic degree. Y-axis is the probability density function for the kernel density estimation. For the mobility network, we estimate the degree of a node as the total number of hourly movements starting or ending in this location across 168 hours in the whole week, as the density plot colored by blue. In contrast, we show the degree distribution of the static mobility network with zones as nodes and passenger flows between nodes as edges, aggregating 2,338,576 trips by taxi passengers in 13,798 taxis in Shenzhen from 18 April 2011 to 26 April 2011 over 1634 zones7, as the density plot colored by black. We fit the two datasets by Gamma distributions for our released dataset and Shenzhen. More details of fitness summaries are shown by texts associated with each plot.