A Synthetic Dataset Simulating PDO Feta Cheese Supply Chains for Fraud Detection Research

This collection provides synthetic datasets representing operational data from a Protected Designation of Origin (PDO) Feta cheese supply chain. The main datasets simulate supplier interactions, transportation logistics, milk collection, and physicochemical quality parameters, while incorporating realistic fraud scenarios such as water dilution and species adulteration. Four yearly datasets (2021–2024) capture individual annual records, and a merged dataset combines all years for longitudinal analyses.

Additionally, derivative datasets are provided for specific analytical tasks: one summarizes supplier-level statistics for clustering and risk profiling, and another aggregates goat milk records with features suitable for supervised fraud prediction. These datasets support the development, benchmarking, and validation of machine learning and analytical models for fraud detection, risk assessment, and behavioral analysis in dairy supply chains.