UN-Knowledge-Graph.pdf

"This presentation serves to empower users in demonstrating an approach on how to get started working with the UN datasets for the process of scoping and modelling data into a knowledge graph. This walkthrough tutorial will focus on how to break down key questions and data into concepts, relationships, and the properties of a knowledge graph with respect to the Sustainable Development Goals (specifically poverty measurements) and linking to news articles."

You need to be a member of The NIC Collaboration Hub to add comments!

Join The NIC Collaboration Hub

Replies

  • I thought creating a Knowledge Graph was a significant endeavor because it required access to data, significant domain and Machine Learning expertise, as well as appropriate technical infrastructure, this makes it look much easier. 

    • Hi Daniel, I think it's just a data modeling exercise.  The Machine Learning stuff is only needed for sophisticated applications that would be harder to do without Linked Data.  FHIR uses an ontology, and the APIs reference the data types in that ontology, so it needn't be a big deal.  What you gain from this is data portability.  For example, since a FHIR patient has a specific definition, and if you send some FHIR patient info as Linked Data, that specific definition sticks with it.  The records can stand alone, and you don't have to agree on a common database schema between vendor systems (because the web ontology is the data definition). As far as storing the records, anything that can store JSON or XML can store Linked Data, or you can split up the Linked Data into fields in a regular DB record.  Say you have some Comma Separated Value file.  You can turn that into Linked Data using a methodology like this, and then it's self-describing (links to each CSV field's web definitions are in the Linked Data's JSON-LD or RDF) and you needn't include a human readable data dictionary to be sent along with the CSV. Hope this helps.  

      Converting CSV to RDF with Tarql
      Quick and easy and, if you like, streaming.
This reply was deleted.