Benchmark Dataset for Blockchain-based Maritime Supply Chain Management System: Cargo Dispute Prediction and Risk Profiling

Maritime transport accounts for nearly 80% of international commerce, generating annual revenue exceeding $15 trillion. Nonetheless, blockchain research in this domain is significantly constrained by the lack of real-world test data, due to business secrecy and competitive sensitivity. Evaluating and authenticating blockchain-based maritime supply chain systems need realistic datasets that accurately reflect the intricacies of global trade operations. In this paper, we present a comprehensive benchmark dataset for building and validating a blockchain-based maritime supply chain management system. The MBSC (Maritime Blockchain Supply Chain) dataset is constructed using a parameter-driven Monte Carlo simulation framework and calibrated against a variety of reliable international sources, including the United Nations ComTrade database, the World Bank’s Logistics Performance Index 2025, the UNCTAD Maritime Transport Review 2025, and Lloyd’s List Intelligence. This dataset comprises 36 international ports across six geographic regions, with a combined cargo-handling capacity of over 319 million TEUs (Twenty-foot Equivalent Unit). It also includes seven major trade routes and 12 commodity categories, organised by actual traffic volume, and together they account for 94% of global maritime trade. Four record types are generated: cargo shipments with 32 attributes following a deterministic finite-automaton lifecycle model, entity registrations for supply chain stakeholders, payment settlements, and dispute cases calibrated to International Chamber of Commerce (ICC) dispute-resolution statistics. The statistical formats are domain-calibrated: log-normal for cargo value, bounded normal for container weight, gamma for transit time, and exponential for customs processing. A central feature of MBSC is the generation of dispute and risk-related variables. Dispute labels are produced using cumulative risk factors, including cargo value, hazardous or perishable cargo status, transit duration, customs delay, documentation issues, and payment conditions. These variables allow the dataset to be reused for binary dispute prediction, cargo risk scoring, and risk-aware decision support. It is designed for reuse in blockchain consensus algorithm development, smart contract testing, dispute prediction modelling, and cargo risk profiling. The dataset is particularly suited for researchers working on maritime blockchain applications who require realistic, openly accessible training data for machine learning experiments, blockchain prototype testing, algorithm development, and performance benchmarking. Files The dataset contains the following files: cargo_dataset.csv: Primary cargo shipment records (1K/10K/100K/1000k variants) dispute_dataset.csv: Cargo dispute records with resolution outcomes entity_dataset.csv: Blockchain network participants (ports, customs, agencies) payment_dataset.csv: Escrow payment transactions ports_reference.csv: Port metadata with geographic and efficiency data figures/: Visualizations Main research uses MBSC can be used for: dispute prediction cargo risk profiling maritime supply chain analytics smart contract workflow testing blockchain-based cargo tracking experiments payment settlement and escrow workflow studies decentralised application prototyping reproducible benchmarking Data generation The dataset was generated through simulation. Publicly available maritime, logistics, trade and arbitration-related sources were used to guide the design where possible. The simulation includes cargo characteristics, route information, customs and duty status, payment records, dispute outcomes and blockchain-style identifiers. Dispute records were created using a risk-based approach. Factors such as cargo value, hazardous or perishable cargo status, transit duration, customs delay, documentation issues and payment conditions were used to shape dispute likelihood. Important note MBSC is not real shipping data. It should not be used to make claims about actual dispute rates, real port performance, company behaviour or operational outcomes. It is intended only for research, testing and benchmarking.