The extraordinary growth in complex data has left many enterprises struggling to create an integrated, comprehensive view of that data.
In recent years, knowledge graphs have emerged as a powerful tool for integrating large volumes of distributed, data, both structured and unstructured.
A simplified graph data model available in graph databases helps enterprises achieve this goal in fewer steps and without locking into a single notion of the analytics needed.
DBTA recently held a webinar with Steve Sarsfield, VP product, AnzoGraph DB, who discussed the powerful side benefits of graph databases, like graph algorithms, and how they go beyond standard analytics to uncover relationships in the data.
A knowledge graph can mean different things to different people, Sarsfield explained. For executives, they see a common understanding of all disparate data while data architects see knowledge graphs as one method to integrate data from multiple data sets, structured or unstructured, and to leverage standard industry ontologies to enhance analytics. And ontologists believe knowledge graphs are the best way to represent knowledge and meaning and provide linkage and relationship information in a data analytics platform. Ontologies are at the center providing a way to standardize and enhance the conceptual model. Inferencing provides semantic reasoning for better understanding.
To build a scalable knowledge graph, Sarsfield said, companies must leverage diverse data sets, perform analytics, and scale.
To leverage diverse data, enterprises can use knowledge graphs with square tables or use a flexible data model with RDF where everything is a triple, Sarsfield said. Ontologies in RDF provide context, classification, equivalencies.
An archived on-demand replay of this webinar is available here.