Building Knowledge Graphs

Preface

Graph databases and graph data science have reached a significant level of adoption. They have been extensively used for a range of discrete use cases like logistics, recommendations, and fraud detection. But there is a bigger emerging trend to arrange data in a deliberate manner that enables insight at scale across functional silos. The technology underpinning this trend is know as a knowledge graph.

The forces behind the trend are clear: organizations are no longer suffering from data scarcity. In fact, in an era when big data seems to be a solved problem (at least from a storage point of view), many organizations are practically drowning in data. Industry anecdotes of many thousands of relational tables per day being ingested into a data lake abound, but with an abundance of data there comes the unexpected challenge of what to with it. This is where knowledge graphs help.

A knowledge graph is a purposeful arrangement of data such that information is put in context and insight is readily available. Individual records are placed in an associative network of relationships that provide rich semantic connectivity and context. That network of relationships—​a graph—​is an incredibly intuitive way of representing useful knowledge. Data that might have originally existed to serve a fraud-detection use case can be repurposed seamlessly within the knowledge graph to provide data for recommending financial products (or vice versa). And from there it is straightforward to connect ...