SN SciGraph data model aims at being comprehensive and expressive, yet simple and intuitive. Its main goal (in accordance with the LOUD principles) is to make it easier for people to understand and reuse our datasets within other projects.

The data model is encoded using the Resource Description Framework language which is used for representing Linked Data on the web. RDF is an abstract model which allows for number of serializations, e.g. Turtle, N-Triples, RDF-XML, JSON-LD etc..

SN SciGraph uses JSON-LD as its canonical serialization. JSON-LD is a fully compliant JSON format, meaning that data can be processed using all the standard JSON tools and libraries, even without any prior knowledge of Linked Data and RDF.


The SN SciGraph model follows the specifications.

There are 8 main classes, only one of which (Patent) required the creation of an ad-hoc entity type (by subclassing schema:CreativeWork) as it did not exist in The remaining seven follow closely the vocabulary and semantics specified on the site.

  • schema:ScholarlyArticle is used to describe journal articles;
  • schema:Chapter is used to describe book chapters;
  • schema:Book is used to describe books;
  • schema:Periodical is used to describe journals;
  • schema:Person is used to describe researchers (e.g. authors, editors, grant recipients)
  • schema:MonetaryGrant is used to describe awarded research grants;
  • schema:MedicalStudy is used to describe clinical trials;
  • sgo:Patent is used to describe patents.
  • For more examples, datasets statistics and identifiers information, see the Datasets: At a Glance page.

    Object Model

    SciGraph Data Model



    Namespaces and other specifications are represented using SciGraph's default JSON-LD context.