Business Intelligence (BI) systems are extensively used as in-house solutions to support decision-making in organizations. Next generation BI 2.0 systems claim for expanding the use of BI solutions to external data sources and assisting the user in conducting data analysis. In this context, the Analytical Metadata (AM) framework defines the metadata artifacts (e.g., schema and queries) that are exploited for user assistance purposes. As such artifacts are typically handled in ad-hoc and system specific manners, BI 2.0 argues for a flexible solution supporting metadata exploration across different systems. In this paper, we focus on the AM modeling. We propose SM4AM, an RDF-based Semantic Metamodel for AM. On the one hand, we claim for ontological metamodeling as the proper solution, instead of a fixed universal model, due to (meta)data models heterogeneity in BI 2.0. On the other hand, RDF provides means for facilitating defining and sharing flexible metadata representations. Furthermore, we provide a method to instantiate our metamodel. Finally, we present a real-world case study and discuss how SM4AM, specially the schema and query artifacts, can help traversing different models instantiating our metamodel and enabling innovative means to explore external repositories in what we call metamodel-driven (meta)data exploration.
- Business intelligence, Metadata, Ontological metamodeling
Varga, J., Romero, O., Pedersen, T. B., & Thomsen, C. (2018). Analytical metadata modeling for next generation BI systems. Journal of Systems and Software, 144, 240 - 254. https://doi.org/10.1016/j.jss.2018.06.039