Enhancing Supply Chain Transparency with Blockchain: A Data-Driven Analysis of Distributed Ledger Applications

Blockchain technology is increasingly redefining supply chain management paradigms with unprecedented levels of transparency, traceability, and trust in the USA. With increasingly complex supply networks worldwide, the integrity and real-time visibility of transactional information become vital for operational reliability and adherence. This study presents a data-driven examination of the ways distributed ledger technology (DLT), specifically blockchain, facilitates increased supply chain transparency across stakeholders through immutable record-keeping and verifiable sharing of data. The main goal of the current research was to create a synthesis of the secure, immutable nature of blockchain and the predictive and diagnostic power of machine learning (ML) to boost supply chain transparency. The dataset used in this work is formatted blockchain logs, extracted from a permissioned, distributed ledger system simulating a U.S.-based supply chain network. Every log entry stores transactional metadata, high-value data such as accurate timestamps of transactions, cryptographic verdicts, digital handovers between supply chain entities (suppliers, logistics providers, distributors), and route signatures, derived from geolocation-based smart contract activators. In the selection of suitable machine learning models, three classifiers that considered the multi-dimensionality of blockchain supply chain data were used. The training and validation approaches were tailored to maintain the models' robustness and generalizability. The dataset was divided into a 70/30 train-test split using stratified sampling to preserve the proportion of fraudulent versus non-fraudulent instances, guaranteeing that both subsets contained a balanced representation of the classes. By looking at the comparative bar plots of the performance of our models on our blockchain-based supply chain dataset, we observed that the Random Forest Classifier had a slightly greater accuracy and F1-score than the Logistic Regression and the XG-Boost Classifier. In the Food and Agriculture industry, supply chain analytics with blockchain technology can greatly improve traceability, specifically under United States Department of Agriculture (USDA) standards. At U.S. Customs and Border Protection (CBP) checkpoints and international borders, blockchain solutions bring significant advancements in verification speed and counterfeit prevention. By applying analytical tools against the recorded events and metadata, organizations in the USA not only track assets and events but also proactively discover potential risks, streamline processes, and gain a greater insight into their supply chain dynamics. Towards the future, some promising avenues of research open up with the combination of blockchain and machine learning. One such exciting area is the blending of smart contracts with automated responses. Lastly, federated learning among decentralized blockchain nodes is a pioneering line of research that might resolve the issues of sparsity and generalizability of the data and avoid the compromise of the decentralized nature of blockchain.