Regional hubs identification in container shipping network based on community connectivity

As a complex combination of linear shipping among ports, the maritime network of container shipping reflects the regionalization and globalization of international trade. It is demonstrated that community structure is detected in the global shipping network, and each community refers to a clustering region in which the port nodes are closely located and trade-connected. The regional transshipment ports that act as hubs among different communities determine the shipping accessibility of inter-community and play an indispensable role in regional trade, while they have not been recognized sufficiently so far. This paper proposes a novel approach to identify regional hubs among port communities based on community connectivity. We first model the global container shipping network in P-space according to the data from the international shipping consulting agency Alphaliner and then detect 15 port communities by employing a modularity-based Louvain algorithm. The neighborhood of each node in the community is analyzed, and the regional hubs bridging the local community and other communities are recognized based on community connectivity. In contrast to most centrality-based approaches, the proposed approach can identify some small regional transshipment ports that occupy an indispensable hub position and have a critical influence on regional shipping trade. The simulation results show that the identified regional hubs need sufficient attention since they significantly affect system robustness in the intentional attack strategy. Some policy implications are provided for relevant decision-makers such as Shipping companies and port authorities.