Json schema

The latest version of the json schema is available at
For a visual and interactive view of the schema, we suggest to use a json schema viewer like (you just need to copy the schema and then you can easily navigate through nodes).

  • Drawing of the schema/data model
  • data model
    Brief description for each and for the non trivial cases, the processes that affect its value
  • the title of a publication comes as is from the source. No need to declare that anywhere
  • the funder of the publication comes either from the source or is inferred. This we must document
  • the refereed field is constructed with some methodology. This we must document

Dump data model overview

Table of main entities

# Entity type Sub-types Description
1 Result Results are intended as digital objects, described by metadata, resulting from a scientific process
1.1 Publication Publications includes all digital research artefacts whose intended use is narrative storytelling of a research activity and its results. Examples are scientific articles, reports, slides, data papers, etc. Although there are exceptions, as each scientist has a large degree of freedom in publishing and interlinking his artefacts, it can be generally assumed that literature artefacts are published with narrative intent. For those specific cases where literature is intended for different use, we in general do not expect scientists to publish such artefacts as literature artefacts. For example when an article is a carrier of readable datasets (e.g. articles with tables) the article is often deposited a second time in a data repository, assigned a new DOI, and marked as a dataset of type “textual”; in the case articles full-texts are used for natural language processing (NLP), scientists will likely create a dataset of type “collection of articles”.
1.2 Dataset include digital research artefacts encoding experimental or real-world observations/measures (e.g. primary data), secondary data derived from programmatic processing of other datasets, or more generally digital representations of facts to be interpreted by a program. The definition is cross-discipline, hence spans across multiple interpretations of datasets, where typologies and granularity obey to different scientific facets. Examples include, but are not limited to: databases (e.g. Worms), records of databases (e.g. proteins in the UniProt database), table files, queries over databases (time-series slices, geospatial maps, SQL queries), media (e.g. images, videos) or collections of media.
1.3 Software Software ​entities represent research software, i.e. software that is an output of research activity. Examples include, but are not limited to: code scripts, web services, and web applications.
1.4 Other Research Product Other research products ​include any research output that is not literature, data, or software. Examples include, but are not limited to: algorithms, scientific workflows/pipelines, protocols, standard operating procedure (SOP), simulations, mathematical and statistical models, but also research packages. Research packages can group a set of research artefacts, but can also include the encoding of a composition logic that binds them together. For example, an instance of a workflow is a package that describes the combination of specific artefacts to implement a scientific process, execute an experiment, etc.
2 Data source OpenAIRE entity instances are created out of data collected from various data sources of different kinds, such as publication repositories, dataset archives, CRIS systems, funder databases, etc. Data sources export information packages (e.g., XML records, HTTP responses, RDF data, JSON) that may contain information on one or more of such entities and possibly relationships between them. For example, a metadata record about a project carries information for the creation of a Project entity and its participants (as Organization entities). It is important, once each piece of information is extracted from such packages and inserted into the OpenAIRE information space as an entity, for such pieces to keep provenance information relative to the originating data source. This is to give visibility to the data source, but also to enable the reconstruction of the very same piece of information if problems arise.
3 Organization Organizations include companies, research centers or institutions involved as project partners or as responsible of operating data sources. Information about organizations are collected from funder databases like CORDA, registries of data sources like OpenDOAR and re3Data, and CRIS systems, as being related to projects or data sources.
4 Project Of crucial interest to OpenAIRE is also the identification of the funders (e.g. European Commission, WellcomeTrust, FCT Portugal, NWO The Netherlands) that co-funded the projects that have led to a given result. Projects are characterized by a list of funding streams (e.g. FP7, H2020 for the EC), which identify the strands of fundings. Funding streams can be nested to form a tree of sub-funding streams.
5 Community/Initiative Communities/Initiatives are intended as groups of people with a common research intent and can be of two types: ​research initiatives or ​research communities​.
1. Research initiatives are intended to capture a view of the information space that is "research impact"-oriented, i.e. all products generated due to my research initiative;
2. Research communities the latter “research activity” oriented, i.e. all products that may be of interest or related to my research initiative.
For example, the organizations supporting a research infrastructure fall in the first category, while the researchers involved in a discipline fall in the second.

Table of the relationships

A relationship in the graph is represented by the following data type, which aims to model a directed edge between two nodes, providing information about the semantic of the relation, its provenance and validation.

field name cardinality type description
1 source ONE Node Represents the source node in the relation
2 target ONE Node Represents the target node in the relation
3 reltype ONE RelType Represent the semantics of the relation between two nodes of the graph
4 provenance ONE Provenance Indicates the process that produced (or provided) the information
5 validated ONE boolean Indicates weather or not the relation was validated
6 validationDate ONE string Indicates the validation date of the relation - applies only when the validated flag is set to true


The Node data type contains the minimum information needed to identify a graph node, its identifier and entity type.

field name cardinality type description
1 id ONE string OpenAIRE identifier of the node in the graph
2 type ONE string graph node type


The RelType data type models the semantic of the relationship among two nodes.

field name cardinality type description
1 type ONE string relation category, e.g. affiliation, citation, see table Relation typologies
2 name ONE string further specifies the relation semantic, indicating the relation direction, e.g. Cites, isCitedBy

Relation typologies

The following table lists all the possible relation semantics found in the graph dump.

# source entity type target entity type relType.type (inverse)
1 Project Result outcome produces isProducedBy
2 Result Organization affiliation hasAuthorInstitution isAuthorInstitutionOf
3 Result Result similarity IsAmongTopNSimilarDocuments HasAmongTopNSimilarDocuments
4 Project Organization participation isParticipant hasParticipant
5 Result Result supplement IsSupplementTo IsSupplementedBy
6 Result Result relationship IsRelatedTo IsRelatedTo
7 Data_source Organization provision provides isProvidedBy
8 Result Data_source provision IsHostedBy hosts
9 Result Data_source provision IsProvidedBy provides
10 Result CommunityInitiative relationship IsRelatedTo IsRelatedTo
11 Organization CommunityInitiative relationship IsRelatedTo IsRelatedTo
12 Data_source CommunityInitiative relationship IsRelatedTo IsRelatedTo
13 Project CommunityInitiative relationship IsRelatedTo IsRelatedTo

Further releases will extend the set of relationship types exported in the graph dump. The candidate relationships are indicated in the following table:

# source entity type target entity type relType.type (inverse)
1 Result Result relationship IsReferencedBy References
2 Result Result citation Cites IsCitedBy
3 Result Result part HasPart IsPartOf
4 Result Result version IsPreviousVersionOf IsNewVersionOf
5 Result Result relationship Continues IsContinuedBy
6 Result Result version IsVersionOf HasVersion
7 Result Result relationship IsIdenticalTo IsIdenticalTo
8 Result Result relationship Documents IsDocumentedBy
9 Result Result relationship IsDerivedFrom IsSourceOf
10 Result Result version IsOriginalFormOf IsVariantFormOf
11 Result Result version Obsoletes IsObsoletedBy
12 Result Result review Reviews IsReviewedBy
13 Result Result relationship Compiles IsCompiledBy


The Provenance data type indicates the process that produced (or provided) the information, and the trust associated to the information.

field name cardinality type description
1 provenance ONE string provenance, contains values defined according to the dnet:provenanceAction vocabulary
2 trust ONE string trust, expressed as a number in the range [0-1] indicates the trustworthiness of the information.

Updated by Miriam Baglioni 5 months ago · 24 revisions