The ability for knowledge graphs to amass information and relationships and connect facts is showing potential for a range of use cases.
Bob Kasenchak, director of business development, Access Innovations, Inc., USA, discussed the rise of knowledge graphs in his presentation, “From Structured Text to Knowledge Graphs: Creating RDF Triples From Published Scholarly Data” at Data Summit 2019.
“Graphs, I think it’s clear, that they’re pretty hot right now,” Kasenchak said.
For scholarly publishers and content driven organizations the content is the most important data asset. It’s commonly expressed in an XML format, Kasenchak explained.
XML is transmittal and is easy for display, but it’s not a queryable data asset. Since the data in XML is fielded it is great to transfer into a knowledge graph.
“Graphs are an excellent way to store and query information that’s full of relationships that can’t be explicated in additional databases,” Kasenchak said. “The goal is to be able to ask questions about data and draw inferences about the content, RDF is an excellent candidate because that’s what it’s designed for.”
Since data is fielded it can extracted from XML and put it in RDF Triples to create that knowledge graph, he explained.
Structured data is abstracted, organized into RDF statements using some existing or custom built ontology, loaded into some repository, and then it’s serviced using an endpoint, according to Kasenchak.
However, even following these steps, there are still pitfalls. Data needs to be clean and large data sets need to be organized before being modeled.
Many Data Summit 2019 presentations are available for review at https://www.dbta.com/DataSummit/2019/Presentations.aspx.