Data analysis for more accurate cargo ship ETA’s: a model for ETA deviation prediction

During the past decades, the focus in maritime supply chains has shifted from the individual company to an integrated chain. It is of utmost importance that the various chain actors are well aligned with each other, so as to make transport operations as smooth as possible, ideally also without Estimated Time of Arrival (ETA) deviations. To examine which parameters have to be tackled primarily so as to reduce ETA deviations, this paper examines for each individual ship type at the port of Antwerp which parameters significantly influence ETA deviations. First and foremost, it is observed that seagoing vessels exhibit significantly greater deviations than inland vessels. We also find that chemtankers do not appear to be significantly influenced by their dimensions and geographic parameters. Container ships, on the other hand, are insensitive to both geographical and meteorological parameters. General tankers, on the other hand, show the greatest deviations. The output of the synthesizing Principal Component Analysis indicates that for all ship types the dimensions and meteorological parameters have the greatest explanatory power in the context of deviation predictions. The equations developed on this basis are able to calculate the chance of ETA deviations for each vessel type in the dataset. This result is therefore of great relevance to all actors in the maritime supply chains, as it allows planning much better and avoiding unnecessary asset and staff expenses.