KnowGraph-MDM: A Methodology for Knowledge-Graph-based Master Data Management

Abstract

In highly globalized, digitized, and complex Supply Chains (SCs), the view of SCs is based on seemingly siloed or outdated data-sets. Integrated analysis capabilities, that rely on consistent data structures and applications, facilitate grasping, controlling, and ultimately enhancing SC behavior. Master Data (MD) is defined as the high-value core information of an enterprise, shared across the business processes and systems of an organization. Master Data Management (MDM) is an essential prerequisite for companies to make agile reporting and correct decisions. Traditional MDM approaches are limited in integrating enterprise information as well as meeting requirements, e.g., stakeholders' involvement for MD analysis and reporting. Therefore, we propose a methodology for a knowledge-graph-based MDM, which relies on establishing a knowledge-graph (KG) layer for building a common understanding of the key business entities and semantic mappings from and to the original data sources. KnowGraph-MDM (KG-MDM) relies on iterations to incorporate stakeholders' inputs, allowing evolutionary development of the MD model. Thus, the ingestion and adoption of the new model increases among the stakeholders via deployment in the organization. We apply the proposed approach in a use case for a semiconductor manufacturer. The resulting KG depicts the core MD model that can be iteratively extended to incorporate different stakeholders' perspectives. KnowGraph-MDM enables integrated SC performance analysis and reporting, relying on consistent MD across the SC.