GRAPHIA’s SSH Federation: The GoTriple Knowledge Graph Goes Live
Authors: Julien Homo, Kévin Darty and Ursula Rabar
The SSH Knowledge Graph that GRAPHIA is building is a federation: a set of autonomous knowledge graphs, each keeping its own governance and data, connected through a common interoperability layer. As such, a federation takes time to build, node by node. Today we are happy to announce that its first fully implemented node is live: the GoTriple Knowledge Graph, and anyone can query it.
As we explained in our previous post, GoTriple, the OPERAS discovery platform for the SSH, is the largest SSH-only repository of scholarly metadata, but its data layer rests on textual indexes rather than Resource Description Framework, so it could not join a federation of knowledge graphs as it stood. The GoTriple Knowledge Graph is the component that changes that: without modifying GoTriple itself, it expresses its content in a common Resource Description Framework model, making it visible and queryable within the federated SSH Knowledge Graph, and opening it to every other graph that joins.
What is Released
The GoTriple Knowledge Graph expresses the GoTriple corpus, publications and author profiles from more than 1,500 SSH sources, as native Resource Description Framework: a graph you can actually traverse, rather than a search index you can only browse. Documents are connected to their authors, disciplines, languages and identifiers through the TRIPLE Ontology, so questions that were previously out of reach become a single query: who co-publishes with whom, how a discipline evolves over time, which publications carry a DOI or an ORCID.
Two public entry points are available:
- kg-api.gotriple.eu – A standard SPARQL 1.1 endpoint with its interactive API documentation, for researchers and developers who want to plug the graph into their own tools. GRAPHIA aims to address one of the major challenges facing SSH research today: the fragmentation of scholarly information across repositories, infrastructures, and national initiatives. Through semantic technologies, knowledge graphs, and AI, the project will make research outputs more discoverable, machine-readable, and reusable.
- kg.gotriple.eu – A web interface where you can explore the graph without any setup. It ships with a library of 27 curated example queries, from “show me a random publication” to co-authorship statistics, a user guide written for newcomers, and live statistics on the corpus.
Everything is served read-only from production infrastructure, powered by QLever, the high-performance open-source engine we selected after benchmarking candidates on a 155-million-triple dataset.
Tell Us What You Think
This release is also a milestone in the project’s open process. The full story, the architecture, the data model, the ingestion chain and what comes next, is told in the report First SSH KG release, published on Zenodo, and we are collecting comments through a short feedback form until the 30th of September. Whether you are an SSH researcher, a data provider or simply curious, we would genuinely like to hear what works, what does not, and what you would want the graph to answer next. Open the interface, run an example query, and tell us.
What Comes Next
This first release is a foundation. The corpus will grow towards the full breadth of GoTriple and beyond, and the AI-driven enrichments introduced in our previous posts, concepts and entities extracted from the full text of publications, will progressively feed the graph. On the federation side, the Federated Gateway, the component that will let a single question reach across GoTriple, OpenCitations, GESIS, ORKG and the other graphs of the federation, has already reached a first internal prototype and advances on its own track. Node by node, this is the “from fragmentation to federation” direction that runs through all of GRAPHIA’s work, and we will keep sharing progress here as it takes shape.