Providing research graph data in JSON-LD using schema.org

In this position paper, we describe a pilot project that provides Research Graph records to external web services using JSON-LD. The Research Graph database contains a large-scale graph that links research datasets (i.e., data used to support research) to funding records (i.e. grants), publications and researcher records such as ORCID profiles. This database was derived from the work of the Research Data Alliance Working Group on Data Description Registry Interoperability (DDRI), and curated using the Research Data Switchboard open source software. By being available in Linked Data format, the Research Graph database is more accessible to third-party web services over the Internet, which thus opens the opportunity to connect to the rest of the world in the semantic format. The primary purpose of this pilot project is to evaluate the feasibility of converting registry objects in Research Graph to JSON-LD by accessing widely used vocabularies published at Schema.org. In this paper, we provide examples of publications, datasets and grants from international research institutions such as CERN INSPIREHEP, National Computational Infrastructure (NCI) in Australia, and Australian Research Council (ARC). Furthermore, we show how these Research Graph records are made semantically available as Linked Data through using Schema.org. The mapping between Research Graph schema and Schema.org is available on GitHub repository. We also discuss the potential need for an extension to Schema.org vocabulary for scholarly communication.