Spatiotemporal patterns and predictive risk analysis of maritime accidents in UK waters
Date:
Tang, J., Gao, Y., Majumdar, A., Gao, X., Ye, Y., Cheng, T.*, & Wang, X. (2026, August 26-29). Spatiotemporal patterns and predictive risk analysis of maritime accidents in UK waters [Poster Presentation]. The 1st International Conference of Maritime Transportation and Innovation & The 4th International Conference on Transport and Supply Chain Resilience and Digital Transformation (ICMTI2026&TSCRDT2026), Liverpool, UK.
Abstract: This document summarises the implemented cluster-first spatial-severity analysis framework and its preliminary outputs. The aim is to identify maritime accident clusters in UK waters, score accident-level severe-outcome risk, explain accident mechanisms inside each cluster, analyse temporal patterns, and synthesise evidence into a transparent cluster-priority table. The cleaned MAIB dataset contains 3,938 occurrence-level records for 2013-2021, including 228 serious or very serious cases. Spatial clustering identified 68 discrete hotspots from 2,766 casualty-with-a-ship spatial candidates. The final outputs distinguish high-frequency operational port clusters from lower-frequency severe-risk watchlist clusters and provide cluster-level evidence for targeted maritime safety management.