Graph Gurus, Episode 1: Building an Enterprise Knowledge Graph from RDF Data
- Blog >
- Graph Gurus, Episode 1: Building an Enterprise Knowledge Graph from RDF Data
Update: As of May 1, 2020 we have 34 Graph Gurus episodes! Find all of them at www.tigergraph.com/webinars.
This month we kicked off Graph Gurus, our webinar series for developers who are looking to address real-world problems with graph database technology and graph analytics. Our goal is to make you Graph-Smart, so you can see when a graph could benefit you, and to give you a head start on designing your solution. We outline a problem or application and then show how to design a solution, with a working example running on TigerGraph.
Graph Gurus:
- Aimed at Developers and Data Scientists
- Presented by Developers
- Code & Data used in the demo examples are available for download and use with TigerGraph Developer Edition
Episode 1: Building an Enterprise Knowledge Graph from RDF Data
We first defined a knowledge graph as a graph where there is a strong semantic meaning to the relationships between entities, and where users often want to draw inference from looking at the network of relationships. I know, it sounds like every graph. An Enterprise Knowledge Graph is simply a knowledge graph where the knowledge domain is focused on an enterprise.
We then explained why it might be smart to use a property graph like TigerGraph to implement your knowledge graph, rather than an RDF:
- RDF data stores typically use relational joins to handle graph traversal → too slow
- RDF/SparQL weren’t designed for fine-controlled or computationally complex analytics → hard to use
Xinyu Chang, our Solutions Team Leader, then guided us through the full implementation of a knowledge graph using TigerGraph, including:
- Introducing our example dataset from DBpedia
- Defining a Universal Graph Schema, valid for any knowledge graph
- Mapping RDF data into our graph database
- Exploring the graph visually in GraphStudio
- Writing and running GSQL queries to answer questions like: “Given a person, find the top K persons who are most related via common objects”
GSQL Universal Knowledge Graph Schema
You can see a full replay of the webinar on TigerGraph’s YouTube channel.
Then please register to see our next webinar live, and ask questions of our developers!