Replication Data and Code for Evaluating Port Performance After Accounting for Fleet Characteristics Across Maritime Economies

This repository contains the replication data, analytical code, model outputs, figures, and supporting documentation for the study “Evaluating Port Performance After Accounting for Fleet Characteristics Across Maritime Economies.” The study develops a Port–Fleet Compatibility Gap (PFCG), which compares observed median time in port with the time expected from vessel age, vessel size, vessel market, and half-year period. The source dataset was accessed through the Maritime Port Performance Dataset hosted on Kaggle. The dataset compiler reports that the underlying statistics were collected from the UNCTAD Data Center. The original file contained 803 aggregated economy–vessel-market–period records. After documented exclusions and data preparation, the primary analytical sample contained 507 observations covering 22 maritime economies, six vessel markets, and four half-year periods during 2022–2023. The package includes a reproducible Google Colab notebook, processed analytical data, grouped cross-validation outputs, model predictions, economy-level compatibility-gap estimates, bootstrap and sensitivity results, SHAP outputs, supplementary TEU/DWT analyses, figures, tables, a README, and licence information. The PFCG is a condition-adjusted diagnostic measure. It is not a causal efficiency score and should not be interpreted as a ranking of individual ports or terminals.