Restricted De-Identified Industrial Analytical Dataset for Entropy-Aware Bayesian Decision-Making

This restricted-access deposit accompanies the invited manuscript “From Measuring Uncertainty to Acting on It: Entropy-Aware Bayesian Decision-Making for Risk-Sensitive Data Science” (HDSR-2026-102). The deposit contains a de-identified line-level analytical copy of the industrial fermentation dataset used in the manuscript. It preserves all 73,212 source rows, exact duration values in seconds, missingness, chronological ordering, derived shift labels, pseudonymous product-group structure, pseudonymous semi-finished work-order structure, and transparent analysis-inclusion flags. The deposit does not contain product names, stock codes, original product identifiers, original work-order numbers, lot identifiers, equipment or vehicle identifiers, exact timestamps, weight fields, fire or yield fields, or the private pseudonymization key. The same industrial analysis pipeline was executed against the authorized original export and this restricted analytical copy. The public summary outputs and the survival figure were reproduced with exact byte equality. Access is restricted because the underlying operational data are associated with a commercially sensitive production environment. Access requests may be evaluated for editorial review, peer review, or qualified academic research subject to confidentiality and organizational approval.