“Knowledge graph technology is not a figment of a science-fiction imagination; the technology exists and is in use today. Enterprises are currently using knowledge graphs to support a new culture of data management."
- Dean Allemang, Principal Partner, Working Ontologist, LLC
Recently writer and technologist, Dean Allemang (author of Semantic Web for the Working Ontologist) joined forces with Cambridge Semantics’ founders, Sean Martin and Ben Szekely, to write a definitive guide all things knowledge graph. The fruit of their efforts, an eBook titled “The Rise of the Knowledge Graph,” offers anyone building a knowledge graph or just interested in the topic a solid introduction to all aspects of this emerging data management architecture.
Your invited to preview this in this blog – I’ve included the introduction below or you can download the whole ebook here. You can also join the authors at a live event to discuss this topic on April 27 – reserve your spot. Happy reading!
The Rise of the Knowledge Graph: Introduction
While data has always been important to business across industries, in recent years the essential role of data in all aspects of business has become increasingly clear. The availability of data in everyday life— from the ability to find any information on the web in the blink of an eye to the voice-driven support of automated personal assistants—has raised the expectations of what data can deliver for businesses. It is not uncommon for a company leader to say, “Why can’t I have my data at my fingertips, the way Google does it for the web?” This is where a structure called a knowledge graph comes into play.
In this report, you will learn:
- What a knowledge graph is and how it accelerates access to good, understandable data
- What makes a graph representation different from other data representations, and why this is important for managing enterprise, public, and research data
- What it means to represent knowledge in such a way that it can be connected to data, and what technology is available to support that process
- How knowledge graphs can form a data fabric to support other data-intensive tasks, such as machine learning and data analysis
- How a data fabric supports intense data-driven business tasks more robustly than a database or even a data architecture
A knowledge graph is a combination of two things: business data in a graph, and an explicit representation of knowledge. Businesses manage data so that they can understand the connections between their customers, products or services, features, markets, and anything else that impacts the enterprise. A graph represents these connections directly, allowing us to analyze and understand the relationships that drive business forward. Knowledge provides background information such as what kinds of things are important to the company and how they relate to one another. An explicit representation of business knowledge allows different data sets to share a common reference. A knowledge graph combines the business data and the business knowledge to provide a more complete and integrated experience with the organization’s data.
What does a knowledge graph do? To answer that question, let’s consider an example. Knowledge graph technology allows Google to include oral surgeons in a list when you ask for “dentists”; Google manages the data of all businesses, their addresses, and what they do in a graph. The fact that “oral surgeons” are a kind of “dentist” is knowledge that Google combines with this data to present a fully integrated search experience. Knowledge graph technology is essential for achieving this kind of data integration.
A knowledge graph is a combination of two things:
business data in a graph, and an explicit representation of knowledge.
An integrated data experience in the enterprise has eluded data technology for decades, because it is not just a technological problem. The problem also lies in the way enterprise data is governed. In a company, distinct business needs often have their own data sources, resulting in independently managed “silos” of data that have very little interaction. But if the enterprise wants to support innovation and gain insight, it has to adopt a completely different way of thinking about data, and consider it as a resource in itself, independent from any particular application. Data utilization then becomes a process of knitting together data from across the enterprise (sales, product, customers) as well as across the industry (regulations, materials, markets). Continuing the knitting metaphor, we call a data architecture of this sort a data fabric.
This report is written for the various roles that need to work together to weave a data fabric for an enterprise. This includes everyone from data managers, data architects, and strategic decision makers to the data engineers who design and maintain the databases that drive the business. Modern companies also have a variety of data customers, including analysts and data scientists. The data fabric provides these data customers with a much broader resource with which to work. Another key role in driving a data fabric is the business architect, the one who analyzes business processes and figures out who needs to know what to make the processes work correctly. For all the people in these various roles, the data fabric of the business is essential to their everyday activities. Building and maintaining a data fabric is a challenge for any enterprise. The best way to achieve a data fabric is by deploying knowledge graph technology to bring together enterprise data in a meaningful way.