Skip to main content

SKG Interoperability Framework

Connecting Scientific Knowledge Graphs

Research knowledge lives in isolated graph databases - each with their own structure. The SKG-IF defines a shared language so they can exchange data, combine insights, and power discovery across the entire ecosystem.

Today: Isolated silos

Knowledge graphs, CRIS systems, and Research Infrastructure catalogues each model research differently. Exchanging data between them, and with other services, requires custom integrations every time.

With SKG-IF: A federated graph

A shared core model, plus community-driven extensions for specific domains — without breaking compatibility.

The foundation

A shared core of six entity types

Based on the RDA SKG-IF recommendation.
These are the entities every knowledge graph talks about, agreed upon by the RDA SKG-IF Working Group.

Agent

Individuals or organisations involved in the creation, publication, dissemination, curation, or assessment of research products.

Data Source

Services or platforms where research products are stored, preserved, made discoverable, and accessed.

Grant

Funding awarded by a funding body to one or more agents.

Research Product

Research outputs of different types, such as research literature, research data, research software, or other categories.

Topic

Scientific disciplines, subjects, or keywords relevant to a research product and its context.

Venue

Publishing or dissemination channels through which research products are made available (e.g. journals, conferences, platforms).
The OSTrails Contribution

From "Scholarly" to "Scientific"

The core model covers bibliographic metadata. OSTrails extends it to capture how data was generated: the instruments, services, and provenance behind research outputs.

Scholarly (RDA Core)

What existing graphs already share: the bibliographic layer.

Publications-Data-Software Authors Grants Publicaitons Venues Citations

Scientific (Extensions)

What OSTrails adds: how data was actually created and processed.
Instruments Services Provenance Geolocation Indicators

What this enables

Practical applications

Smart DMPs

A DMP tool searches the knowledge graph to automatically fill in details about datasets and grants, reducing manual entry for researchers.

Cross-graph Linking

Link datasets and authors across different graphs and catalogues (including national CRIS) to enrich navigation and discovery.

Impact tracking

Identify publications that cite specific datasets, link research products via identifiers, and track how outputs are reused.

The SKG Commons

Everything you need to implement it

The SKG Commons provide the complete, reusable foundation: core model, extensions, mappings from existing schemas, and a standardised API.

Core Data Model

Six entity types, their attributes and relationships. Implemented as an OWL ontology (SKG-O).

Extensions

Community-driven additions for instruments, services, provenance, geolocation, and more.

Mappings

Translations from DDI, RO-Crate, and OpenAIRE Graph into SKG-IF format.

Common APIs

Standardised OpenAPI interface: lookup by identifier, search with filters, JSON-LD responses.

Explore the full SKG Commons