This research proposes an analytical framework for exploring intra-urban goods movement patterns by integrating spatial analysis, network analysis and spatial interaction analysis. Using daily urban logistics big data (over 10 million orders) provided by the largest online logistics company in Hong Kong (GoGoVan) from 2014 to 2016, we analyzed two spatial characteristics (displacement and direction) of urban goods movement. The origin–destination flows of goods were used to build a spatially embedded network, revealing that Hong Kong became increasingly connected through intra-urban freight movement. Finally, spatial interaction characteristics were revealed using a fitting gravity model.
Workflow of The Empirical Analysis
1. Spatial Characteristics of Goods Movement
We observe that the probability distributions for datasets of different years displayed similar trends and could be well fitted by a bimodal Weibull distribution. Specifically, two maximum points exist, namely, 6 km and 22 km. This means that the count of goods movement did not decrease monotonously with distance between origin and destination. Interesting to note is that the intra-urban freight movement displacement distributions failed to follow an exponential or power law distribution. According to the depiction in subsection on the measurement of spatial characteristics, we can conclude that average freight distance gradually increased. We can also conjecture that increasingly improved transportation networks provided greater convenience for freight, especially long-distance freight.
Displacement distributions in different years: (a) probability distributions for the datasets; (b) cumulative probability distributions for the datasets.
Direction distributions over 3 years
2. Characteristics of the spatially embedded network of goods movement
We constructed a spatially embedded network based on origin–destination flow between units and conducted network analysis. By comparing the network properties for the networks in the years 2014, 2015 and 2016, we observed that Hong Kong became increasingly connected from the perspective of logistics. Furthermore, we investigated the distributions of degree and strength of nodes and examined the correlation between them. We found that their relationships could be well fitted by exponential functions, and all values of goodness-of-fit R2 reached 0.72 or higher. In other words, the freight flows between subdistricts could be estimated by the connectivity of subdistricts at the aggregate level.
Freight movement flows for network construction
Cumulative probability distributions and correlations of degree and strength
3. Characteristics of Spatial interaction
We explored the spatial interaction characteristics of intra-urban freight movement and how the interaction flows were related both to the population (or total trips) of the origin and destination and to the distance between subdistricts by fitting the gravity model. The significant linear relationships between interaction flow and the product of the subdistrict populationsPiPj were observed by fitting the general form of the gravity model. In addition, we found that gravity was suitable for predicting goods movement flows by comparing the estimated interaction flows and observed interaction flows. However, the distance decay parameterβ was significantly smaller than that of human mobility patterns. We concluded that the spatial interaction of goods movement was not substantially influenced by the distance between origin and destination.
Comparison of estimated interaction flows with the observed interaction flow based on the datasets for the years 2014, 2015 and 2016
REFERENCE
Zhao, P., Liu, X., Shi, W., Jia, T., Li, W., & Chen, M. (2020). An empirical study on the intra-urban goods movement patterns using logistics big data. International Journal of Geographical Information Science, 34(6), 1089-1116.
Contact: Dr. Xintao Liu (xintao.liu@polyu.edu.hk)
