An ecosystem of open linked research data creates new opportunities. Enhance your services and tools by leveraging the Connectome Knowledge Graph for intelligent linking of metadata.

The Research Data Connectome 

Linked Data Pipeline to harmonise and structure metadata

The Connectome Partnership has built the Linked Data Pipeline to harvest,  harmonise, enrich and link the metadata of open data resources for research coming from varying spaces e.g. institutional repositories, administrations, NGOs, galleries, museums, archives, and libraries. The metadata is structured and linked using the RESCS ontology, which in itself is continuously adapted to the needs of the research communities. 

Connectome Knowledge Graph to store relationships

The Linked Data Pipeline then stores the structured linked open data in the form of the Connectome Knowledge Graph, a data representation most suitable for research, education, and innovation purposes. This graph contains valuable linked research information on research projects, research grants, datasets, publications, organisations and people.

Connectome API to extract information and insights

The Connectome API then allows the creation of a query library aimed to extract specific information from the Connectome Knowledge Graph. In this way, it enables service providers to fully leverage the linked open data from the Connectome Knowledge Graph and to create added value for scholars and citizens. The first application enabled to use the Connectome API is the prototype discovery platform AskReco. Other use cases are going to be implemented throughout 2021, including the enrichment of the DaSCH Metadata Browser and the Swiss AI Research Overview Platform SAIROP

Connectome AI to enrich value extraction

The Connectome Artificial Intelligence represents a set of techniques such as graph statistics, machine learning, and natural language processing to analyse, explore and enrich the linked open data stored in the Connectome Knowledge Graph. The enriched linked open data is fed back into the Connectome Knowledge Graph. This allows all service providers to extract enriched value from the Connectome Knowledge Graph via the Connectome API.  

How could you benefit from linked open data to manage your research information?

Current Key features for using the Knowledge Graph

 

Feature

Current Status

Harvesting of Data from various providers using dumps, APIs, datastreams and OAI.

enabled

Cleaning and enrichment of metadata for the harmonisation process.

enabled

Harmonisation of different metadata from Data Providers based on the  RESCS Ontology.

enabled

Enrichment of the metadata describing the Data Providers by using Wikidata and linling of the the metadata in the Connectome Knowledge graph 

enabled

Author Name Disambiguation using graph embedded Neural Networks to link metadata.

in development

Adding Provenance information to metadata records.

in development

Harvesting, cleaning, harmonisation and linking of archival metadata by connecting the RESCS Ontology to discipline- and / or use-case specific ontologies. 

in planning, first implementations likely in 2021

Feature

Current Status

Query entities  in the Connectome Knowledge Graph built on the  RESCS Ontology

enabled

Authenticate users through OIDC endpoint (such as Edu-Id)

enabled

Fetch full resources of Knowledge Graph and export linked metadata resources in different formats

enabled

Knowledge graph queries for people and projects related to specific keywords.

enabled

Similarity-based knowledge graph queries.

enabled

Knowledge graph queries for relations (e.g. Project X is related to Publication Y of Author Y).

enabled

Feature

Current Status

Discovering most important (central) nodes in the Knowledge Graph to highlight relevance of a given Researcher for a given topic

enabled

Discovering most important nodes in the Knowledge Graph to highlight relevance of a given Researcher as bridge between disciplines

in development

Identify common knowledge between different research fields on the basis of keywords

in development

Named Entity Recognition and text mining in full text of publications to get new keywords, links and other metadata

in planning, first tests until end 2021

Summarize content such as publication abstracts for various Stakeholder needs (e.g. students, school children, the public etc.)

in Planning, first tests until end 2021

Identify possible future research collaborations on the basis of similar keywords in research outputs using Natural Language Processing.

in planning

Automatically detect and extract bibliographic references and citations from the full text publications to enrich metadata and create new links 

in planning

Automatically detect and extract research methods, research questions or other relevant content metadata from the full text publications to enrich metadata and create new links 

in planning

All current Connectome API and Connectome AI features are documented here. Please contact us as mentioned below if you want to access this data.

General Info for Service Providers

 

  • Harmonised, enriched and linked Metadata of the following entities: Publications, Datasets, Research Projects, Organisations, Grants, Persons
  • Provenance information and solution for name disambiguation e.g. by using Neural Networks (Status: in development).
  • The metadata schema , which was defined with academic stakeholders and data providers, is available via rescs.org.
  • We are working on linking discipline specific metadata / schemas as well as GLAM Data Provider (current focus on Archives).
  • As of beginning August 2021 we harvest metadata from SNSF P3, Innosuisse Aramis, opendata.swiss and Unpaywall
  • We are harvesting from OpenAIRE, Europeana for testing purposes.
  • We will add the following data providers until the end of 2021: SWISSUbase, DaSCH, SARI, BASE, CORE, World Bank Open Data, BCUL Patrinum Archive.
  • We enrich our harvested metadata with other metadata e.g. coming from Wikidata and Microsoft Academic Graph, AMiner & OpenAI.
  • We currently harvest metadata using download dumps, API Accesses, SPARQL Endpoints and OAI-PMH Harvester.
  • Future harvesting capabilities such as Web Crawling will be evaluated in 2022.
  • Re(use) Linked Open Data from Connectome Knowledge Graph in their services through various formats.
  • Fetch, analyse, visualise Linked Open Data and their relationships for their services.
  • Gain insights leveraging enriched Linked Open Data (e.g. through Natural Language Processing, Machine Learning, Network Analysis etc.).
  • Re-Use Open Linked data from the Connectome Knowledge graph.
  • Support in the identification of new relevant data providers.
  • Contribute to the Connectome Special Interest Group (SIG) which supports the identification and implementation of new AI based insights on top of the Connectome Knowledge Graph (using Natural Language Processing, Machine Learning, Network Analysis techniques). This SIG will launch in 2022.
  • Co-Design of features for new and existing Open Research Data Services.
  • Training and Consulting on Connectome API & Connectome AI use.
  • Please feel free to reach out via connectome@switch.ch
What about your tools and services?
How could they be enhanced by leveraging the linked open data of the Connectome Knowledge Graph?

 

Please contact us to talk about your problems or
use this template to describe your user story for the Connectome.

 

Questions? Drop us a line

Newsletter

Keep up to date with the latest developments and
sign up to our Research Data Connectome Newsletter.