Structured Expert-Rule Pipelines for Analytics Method Selection in Supply Chain Systems

Selecting an appropriate analytical method for a given supply chain decision problem is a recurring challenge for both researchers and practitioners. Existing taxonomies classify problems and methods independently but offer limited guidance on how data-availability conditions constrain the feasible set of methods. This paper introduces three structured expert-rule pipelines—descriptive, predictive, and prescriptive—that condition on observable data-availability predicates to prescribe a method family for Supply Chain Systems (SCS). Each pipeline is structured as a binary decision tree whose internal nodes are Boolean predicates on data conditions and whose leaf nodes map to concrete method families. The pipelines are designed to be usable by operations managers with limited statistical training, subject to expert interpretation of the predicates and to validation of the resulting outputs. The framework is domain-independent: while it applies to any supply chain decision context, this paper develops its sustainability dimension by grounding all application examples in environmental, social, and governance (ESG) scenarios—including carbon footprint monitoring compliant with the Greenhouse Gas (GHG) Protocol and ISO 14064, energy consumption forecasting for Corporate Sustainability Reporting Directive (CSRD) and European Sustainability Reporting Standards (ESRS) reporting, and green vehicle routing aligned with ISO 14040/44 life-cycle assessment principles—to demonstrate how the same methodological structure supports circular economy analytics, ESG reporting, and decarbonisation objectives. A retrospective application to a proprietary operational distribution dataset with two use cases shows that the pipelines are executable and discriminating: the same physical system routes to different feasible method sets depending on data content and depth, and each unreachable branch is paired with the specific field whose collection would enable it. The paper contributes an expert-rule-based selection tool that bridges the gap between problem taxonomy and practical method deployment.