Introduction to Knowledge Graphs and Key Elements

Knowledge graphs sit behind much of how modern search works, and they are increasingly used to ground generative AI. This lesson opens the Knowledge Graphs primer by explaining what a knowledge graph is, which elements it is built from, and how it relates to the ontologies that often shape it. It gives you the vocabulary and mental model needed for the more technical lessons that follow.

What this lesson covers

• A clear definition of knowledge graphs and their core principles
• Google’s Knowledge Graph and where it appears in search, including knowledge panels
• The key building blocks: nodes, relationships and organising principles
• How knowledge graphs store connected data in graph databases
• What ontologies are and how they differ from knowledge graphs
• Applications in search, discovery, question answering and grounding generative AI

A knowledge graph is described as a structured database that stores information as a graph, connecting entities, their attributes and the relationships between them in a meaningful way. Entities can be people, objects, institutions, events, situations or concepts, and relationships capture the context of how they connect. Because the data is stored consistently, knowledge graphs suit real-time applications, search and discovery, and grounding generative AI for question answering, where they can help reduce hallucinations. The lesson links this to everyday search: Google identifies entities in a query, uses its Knowledge Graph to find related entities, and sometimes shows a knowledge panel that combines what it knows.

The lesson then breaks a graph into its parts. Nodes store details about entities and carry labels for their type. Relationships, also called edges, link two nodes and have their own labels and properties. Organising principles provide the framework that arranges nodes and relationships around the concepts that matter for a use case. A simple e-commerce example helps distinguish instance data from organising principles. One strength highlighted is that knowledge graphs can combine several organising principles, which makes them a good fit for very large collections of facts. The final part explains ontologies as formal specifications of concepts and relationships in a domain. An ontology acts as the blueprint and supports reasoning, while the knowledge graph implements it with real-world data for querying, visualisation and analytics.

This is the first lesson of the optional Introduction to Knowledge Graphs primer, which leads into how Google’s Knowledge Graph works and a practical lesson on its API. It suits SEOs and marketers who want solid foundations before working on the course’s knowledge graph modules.