Digital Governance and Feasibility of a Port Community System (PCS) at the Port of Santos

a data-driven and stakeholder integration approach

Abstract

This article presents a feasibility study for the implementation of a Port Community System (PCS) at the Port of Santos, focusing on the transition from theoretical best practices to quantitative application. The Brazilian port sector exhibits high informational fragmentation, impacting the efficiency of national foreign trade. The methodology is grounded in the Data Science lifecycle, employing an unprecedented integration of datasets from ANTAQ, Porto Sem Papel (PSP), and Regulatory Consent Statistics. Preprocessing techniques such as KNN Imputer, One-hot Encoding, and Standard Scaler were applied, followed by an XGBoost predictive model optimized through the Optuna framework to analyze Lead Time bottlenecks. The results, validated by SHAP values and cluster analysis (K-Means), demonstrate that multistakeholder governance supported by predictive tools is capable of mitigating organizational and technical barriers. The study proposes an integration framework that leverages data analysis to support shared decision-making, aiming to enhance the competitiveness of the Port of Santos complex.