What is a Knowledge Graph? | Yext Hitchhikers Platform

What You’ll Learn

In this section, you will learn:

• What a knowledge graph is
• How a knowledge graph powers consumer experiences on search, business listings, and webpages

What is a Knowledge Graph?

A knowledge graph is a database that is designed to structure data objects and the relationship between those objects. For example, a restaurant franchise (like Galaxy Grill — the fictional brand you’ll see across Hitchhikers) could use a knowledge graph to store data on objects like:

• Restaurant locations
• Menu items
• Services like delivery and takeout
• Job postings

From there, a knowledge graph could help the brand to create relationships between those objects:

• Menu items that are available at certain restaurant locations
• Which locations offer delivery
• Job postings by restaurant location

These object relationships are the foundation of how a knowledge graph — like the Knowledge Graph in Yext — makes brand information available and useful to consumers.

Why Use a Knowledge Graph?

The purpose of a knowledge graph is to power downstream experiences. A downstream experience is any application that can use the data objects and their relationships in order to build a user-facing experience. These can include:

• Business listings on Google, Apple, and other major publishers
• Web pages
• Site search experiences

Because a knowledge graph uses object relationships, that data can be used to answer complex questions from consumers. As an example, let’s look at how data stored in a knowledge graph with object relationships can power a dynamic site search experience.

We’ll use this example search: “Financial advisor in New York who speaks Mandarin”

This seems like a simple search query. However, this query actually contains a few different object relationships — in this case, a financial advisor, a location, and a language.

If this data is not stored in a way that reflects these object relationships, this search query becomes harder to parse. This is an example of how a traditional keyword index search might parse this question:

Traditional index-based search typically doesn’t handle queries like this well, because it doesn’t account for the relationships between the different objects that are in a query.

If this data was stored in a knowledge graph, with object relationships in place, that data could be used to build a semantic search experience. The results may look something like this:

In the next unit, you’ll learn about how objects and relationships are structured in the Yext Knowledge Graph, and how that structured data powers products across the Yext platform.