Explore open social sciences and humanities research from GoTriple as a connected graph. Trace co-authors, topics and disciplines, then share or export your results. .
Run a query on the GoTriple KG
SPARQL query results
SPARQL query
Run a query to see results
User guide
How to query
The page lets you write and run SPARQL queries against the GoTriple knowledge graph.
1. Start from an example
The fastest way is to pick a ready-made query and change it. Select Browse examples. You can search the list, copy a query, or select it to edit it.
2. Write your own
Type SPARQL straight into the editor. Add the prefixes you need, a scope and a LIMIT. Queries without them may be refused.
3. Run it
Select Run query. A query that takes longer than 30 seconds is stopped. If that happens, add a LIMIT or make the pattern narrower.
4. Read the results
Results appear as a data table. When the result is a count, it can also show as a breakdown (bars) or a trend (line over time). Select a row to see everything the graph knows about that resource. Select Explore results in the graph to see connections. A note above the table tells you which values are missing, and where.
Shared connections draws the pairs your query returned, with thicker lines for bigger numbers. If the query also returns two numbers for each item, such as ?docs1 and ?docs2 for the two sides of a pair, the items are placed on axes. If the numbers are almost all the same, the axes are left out and a note says why.
Hide duplicates: the same work can be held by more than one repository, so it may show up several times. By default the table shows each work once and tells you how many other versions are hidden, for example “30 results · 4 duplicates hidden”. Turn Hide duplicates off to see every version as its own row. In the graph, a work with other versions shows a “+2” badge, and the graph has its own Hide duplicates switch. On the results page the switch is only there for the table of resources.
5. Share and export
Your recent searches are kept in this browser, on the home page. Share gives you a link. Export downloads the results as a file.
What is in the graph
Everything in the graph is one of a few kinds of thing. Most queries start from a Resource or an Author and follow the links from there.
Resources
A resource is a publication: an article, a book chapter, a dataset, a thesis, a report and so on. In SPARQL its class is gto:Document.
What it holds
Predicate
Title (one per language)
schema:headline
Abstract (one per language)
schema:abstract
Language
schema:inLanguage
Published and last changed
schema:datePublished, schema:dateModified
Authors, in order
schema:author
Publisher and source it came from
schema:publisher, schema:provider
Free keywords
schema:keywords
Discipline
sioc:topic
Type of content, licence
gto:hasContentType, gto:hasLicense
DOI, ISBN, ISSN and other identifiers
datacite:hasIdentifier
Authors
An author is a person or an organisation credited on at least one resource. Its class is gto:Profile. It has a name (usually “Last, First”), the disciplines it works in, and identifiers.
How resources and authors connect
Authors are stored as an ordered list, because the order of authors matters. Each author sits at a position (rdf:_1, rdf:_2 and so on). If you do not need the order, ignore the position:
Two-letter codes written as plain text, such as fr, en, es, de, it, pt. Anything else is other.
Dates
Dates are stored as text that starts with YYYY-MM-DD. To group by year, take the first four characters:
BIND(SUBSTR(STR(?date), 1, 4) AS ?year)
Use it from code
The graph is open. You can query it from this page, from a script, or from another application. You do not need an account or a key. It is read-only.
Endpoint
Send queries with GET (the query in the address) or POST (the query in the body, or as a form field named query).
Result formats
Choose one with the Accept header.
application/sparql-results+json, …+xml
SELECT and ASK
text/csv, text/tab-separated-values
SELECT
text/turtle, application/rdf+xml
CONSTRUCT and DESCRIBE
Good to know
Requests from other websites are allowed (CORS), so a web page can query the graph directly. The endpoint is rate-limited. Queries that run longer than 30 seconds are stopped.
About GoTriple KG
The GoTriple Knowledge Graph is the linked-data version of everything GoTriple knows about research in the social sciences and humanities. It puts all of GoTriple's records into one graph, and adds new information that AI services extract from the full text of documents. You can query it through a public SPARQL endpoint.
Part of the GRAPHIA knowledge graph
GoTriple KG is the social sciences and humanities hub of a larger network built in the GRAPHIA project. GRAPHIA does not merge everything into one place. Each graph stays independent and they are connected through a shared model. GoTriple is linked to OpenCitations, GESIS and ORKG, among others.
Background
GoTriple is the discovery service of OPERAS, the research infrastructure for open scholarly communication in the social sciences and humanities. It is also a service of the European Open Science Cloud. It was inspired by the Isidore platform of the CNRS and is the main result of the TRIPLE project, funded under Horizon 2020.
Credits and funding
The knowledge graph is built in the GRAPHIA project. Net7 and Foxcub lead its construction. PCSS provides the infrastructure. The data model comes from the TRIPLE project.
Funded by the European Union (grant 101188018). Views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or the granting authority.
Records come from many sources, and each one keeps the licence of its provider, shown in its metadata. Check the licence before you reuse a result.
Select a query to run it or to use it as an editable template.
Predicates in the graph
A predicate is the name of a relationship or property, such as “author” or “title”. Select one to copy its name.
What can I do here?
What this is
Search the GoTriple knowledge graph with SPARQL, then explore what you find.
What you can do
Write a query or start from an example, then scan the results.
See the same results as a table, a chart or a graph.
Explore connections from any result and inspect any entity's details.
Some works appear more than once, for example a copy held by another repository. Hide duplicates shows each work once. It is on by default, and you can turn it off in the table.
Share a query as a link, or export the results as a file. Your recent searches are kept in this browser.
What's searchable
Select a resource to start exploring the graph.
Explore in the graph
Starting from …
How it works
What you are looking at
This view shows the real, stored relationships around one entity. Every line is an actual property in the knowledge graph, labelled with its name. Hover a line to see why two things are connected.
ResourceAuthorDisciplineGroupedOther version
Reading the graph
A dashed outline means the node has more connections you have not expanded yet. Solid means fully expanded.
An amber “N resources” node is a collapsed group, used when a relationship has too many items to draw at once. Select it to reveal a sample.
A larger resource node with a “+2” badge has other versions of the same work (for example a copy held by another repository). With Hide duplicates on, they stay out of the way: select the node and choose other versions to show them.
Light grey nodes on a dashed line are those other versions. The dashed line means the match is a best guess (same title, date and first author), not a link stored in the graph.
How to explore
Select any node to open its menu.
Under Connections, pick a relationship type (each shows a count) to expand just that one. Large ones collapse into a group node.
Details opens the entity's full properties. Expand all expands every relationship. Focus dims everything unrelated. Remove prunes the node.
Controls
Max nodes caps how much is drawn at once. Raise it to keep expanding.
Change starting point lets you pick another resource from your results. Back to results closes the graph.
