Unveiling Causality in Stateful Enterprise Knowledge Graphs: An Exploration of Process-Driven Object Relationships

Schobert M, Mala M, Herhoffer M, Farag S, Franke J (2024)

Publication Type: Conference contribution

Publication year: 2024

Journal

Procedia CIRP Curran

Publisher: Elsevier B.V.

Book Volume: 128

Pages Range: 405-410

Conference Proceedings Title: Procedia CIRP

Event location: Cranfield, GBR

DOI: 10.1016/j.procir.2024.03.020

Abstract

The exponential growth of enterprise data generation demands innovative approaches for efficient information management across highly interconnected domains. Particularly, the process of systems engineering involves inhomogeneous data sets requiring a high degree of flexibility and traceability throughout its lifecycle. While structured data resides comfortably in dedicated systems, managing unstructured information within annotations, documents and emails often comes with challenging holistic comprehension of linking, context and revisioning. Graph technology holds promise in establishing adaptable connections among decentralized data sets. However, leveraging graph databases comes with challenges. Schema standardization, crucial for framing complex processes, is a difficult undertaking due to evolving data interrelationships and changing requirements. This paper proposes a two-fold solution. First, we introduce a conceptual schema that decouples stateful object descriptions from its headers and furthermore documents transactions. This allows flexible object associations, conditional state versioning and improved traceability while maintaining manageability. Second, we leverage the Business Process Model and Notation (BPMN) for process design, service orchestration and automated graph manipulations. Through application to the publicly available Northwind data set using neo4j and Flowable BPM, we illustrate our methodology's efficacy and the seamless integration of type-safer data science into dynamic knowledge graphs, contributing to the future of enterprise data management.

Authors with CRIS profile

Marvin Schobert Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS) Jörg Franke Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS)

Involved external institutions

Flowable Deutschland GmbH DE Germany (DE) exentra GmbH DE Germany (DE) Frends Technology FI Finland (FI)

How to cite

APA:

Schobert, M., Mala, M., Herhoffer, M., Farag, S., & Franke, J. (2024). Unveiling Causality in Stateful Enterprise Knowledge Graphs: An Exploration of Process-Driven Object Relationships. In John Ahmet Erkoyuncu, Maryam Farsi, Pavan Addepalli (Eds.), Procedia CIRP (pp. 405-410). Cranfield, GBR: Elsevier B.V..

MLA:

Schobert, Marvin, et al. "Unveiling Causality in Stateful Enterprise Knowledge Graphs: An Exploration of Process-Driven Object Relationships." Proceedings of the 34th CIRP Design Conference, CIRP 2024, Cranfield, GBR Ed. John Ahmet Erkoyuncu, Maryam Farsi, Pavan Addepalli, Elsevier B.V., 2024. 405-410.

BibTeX: Download