Port selection by container ships: A big AIS data analytics approach

Abstract

Port selection is of vital importance for both port operators and shipping lines. In this contribution, an Automatic Identification System (AIS) big data approach is developed. This approach allows identifying container ships using only AIS data without the need for supplementary information from commercial databases. This approach is applied to investigate the port selection statistics of container ships between Shanghai and Ningbo Zhoushan Port, two of the largest ports in the world in terms of calling frequency, to generate practical insights. Results show that: i) the ratios among large ships, medium ships and small ships of these two ports are both approximately 1: 4: 5; ii) these two ports both have an exclusive (i.e., more feeder ports covered in geographical coverage) and intensive (i.e., more feeder ships deployed in shipping service frequency) collection and distribution network mainly consisting of small ships, but that of Shanghai is more intensive; iii) in terms of ultra-large ships over 380 m, Shanghai has accommodated an extra 18.5% compared to that of Ningbo Zhoushan, this indicates Shanghai's attraction for such vessels in global fleet deployment; iv) the feeder network between Shanghai and Ningbo Zhoushan is weak, and their relationship is actually in competition; v) Ningbo Zhoushan could offer more choices for ultra-large container ships (over 380 m), which implies its greater potential in future port competition; vi) when the depth of channels and berths is sufficient, the distance to hinterland and the convenience of a collection and distribution network begin to get more important in port selection. The empirical findings unveil the decision-making of container lines, competition between ports and implications for shipping policy.