Keynote: Michael Schmidt

Abstract: Organizations that want to build new applications using the relationships in their data are confronted with a choice between RDF and property graph models. Today, this choice may have long-ranging ramifications (e.g., when applications must interoperate with other systems), where instead it would be desirable to have users benefit from the best of each world. Motivated by real use cases in the context of Amazon Neptune, a fully managed graph database service that supports both SPARQL queries for RDF as well as Apache TinkerPop Gremlin queries for property graphs, we will highlight aspects that cause organizations to prefer property graphs over RDF (or vice versa) when building their graph applications. Advocating a “graph first, semantics follows” paradigm, we encourage the community to work towards a unification of existing graph data modeling and management approaches. Alongside, this includes exploring improvements of the Semantic Web technology stack for graph use cases (e.g., advanced path queries, graph analytics, and machine learning) and a stronger focus on its usage in the enterprise context. With the ongoing proliferation of graph databases in the industry, a shift in focus towards interdisciplinary, enterprise grade graph data management research could open up a unique chance for the community to make Semantic Web technologies mainstream, starting out from within the enterprise rather than the Web.