Overview: The rise of predictive systems and AI-enabled automated workflows signifies the growing inclination of organizations towards data-driven methodologies. However, there is still a plethora of different types of data sources in a single corporation, ranging from structured relational databases to sophisticated excel table computations. More often than not, these sources have a loose affiliation with fundamental ERP systems. The range of readily available data-driven tools, often as Python packages and its type-agnostic data frames, gave rise to concepts such as data lakes and data meshes. Historically, we've seen a transition from relational databases to data warehouses characterized by star schemas. Now, we're witnessing the emergence of data meshes, signaling yet another shift in how data is structured to attribute its inherent semantics of business value. This development leads to the question: “How can organizations integrate the data that encapsulates their value-creation process”?
Finding a structure that can unite different applications with how Enterprises create value is a highly complex task often not necessarily addressed by current database and data modeling approaches. Within my research, one of my main research streams addresses how to structure data for organizations that capture the inherent complexities of Enterprise value creation. Business logic and data structure exist in conjunction, yet a unifying formal model that captures their relation was missing. I conceptualized an adapted version of directed hypergraphs as a value graph (previously known as a service graph) to accurately capture the fundamental structure of Enterprise value creation.
Origins in service science: I have dedicated part of my time to understanding how value creation can be modeled within organizations. This started within the field of service science. Service science follows the metatheoretical framework of service-dominant logic for understanding value, where “service” is the fundamental basis of all exchange. Service systems are the basic unit of analysis of service science researchers. Service systems are resource configurations that provision value to any other actor. As such, Enterprises are such actors, and information systems are supporting implementations of such service systems. This exploration led me to a realization: within Enterprise Systems, value creation can be conceptualized as a structure on how different resources are configured to provide value.
Traditionally, two main structures capture value creation: a data structure, often hierarchical, coupled with a process framework that bestows an added dimension to the data structure, generally through input-output correlations. Conventionally, this is modeled using two distinct but logically related conceptual modeling techniques (e.g. ER diagrams or BPMN). However, by relying on an adapted directed hypergraph structure, I am able to capture both process and data structures within a single structure. The resulting dir-hypergraph structure can be used as a foundational structure for other Enterprise systems. I am currently working on finding different application areas while refining the foundational structure of the dir-hypergraph (value graph).