Data integration and disintegration: Managing Springer Nature SciGraph with SHACL and OWL

We give an overview of the technical challenges involved in building a large-scale linked data knowledge graph, with a focus on the processes involving the normalization and control of data both entering and leaving the graph. In particular, we discuss how we are leveraging features of the Shapes Constraint Language (SHACL) [1] to combine closed-world, constrained views over an enterprise data integration setting with the open-world (OWL), unconstrained setting of the global semantic web, as well as providing specific data disintegration subsets for data publishing clients.