One could say that there are two key innovations at the core of knowledge graphs 1) Graph database technology, and 2) the ability to semantically model complex relationships. Organizations from a wide array of industries are discovering the benefits of one or both of these innovations. But the blend is becoming increasingly common and beneficial to enterprises and start-ups alike.
“By 2025, graph technologies will be used in 80% of data and analytics innovations, up from 10% in 2021, facilitating rapid decision making across the enterprise.” - Gartner “Market Guide: Graph Database Management Solutions
“Graph techniques are a key component to modern data and analytics capabilities because they span linguistic and numeric domains. Data and analytics leaders must first adopt graph technology and then promote the value it adds in answering increasingly complex business questions.” (How Graph Techniques Deliver Business Value Gartner Peer & Practitioner Research 3 March 2022)
There are many reasons for data and analytics leaders to embrace graph techniques. These techniques can provide unique insight into business problems, especially those problems requiring contextual awareness of the connections and disconnections between multiple entities, including organizations, people, transactions, and events. A well-designed and richly populated graph can capture essential data relationships and their variable nature, including:
- Where data resides
- How various data points may be related
- Who uses the data, and why, when, and how they use it
This article will present use cases that go beyond the basic graph database, as they can be transformative to industry leaders. As they consider which graph database management solution to adopt, these leaders should look to the future and consider expanding the graph projects and ontology/taxonomy projects already in play at their enterprises.
Healthcare & Life Sciences
The healthcare and life sciences industries adopted graph technologies early, taking advantage of their ability to capture essential data relationships and show contextual awareness.
Putting high-quality data into the hands of experts is critical. If an enterprise can unify data from any source, it can:
- Uncover insights from research and development data to accelerate drug discovery
- Put actionable data at researchers’ fingertips
- Make precision medicine a reality
- Gain insight into public health
Drug discovery is a popular graph database use case. Bioinformaticians and analysts need to access all their institutional knowledge—and often data found outside their institute. This means they must connect data from disparate parts of the company to increase research and operational efficiency. This approach allows them to increase output and ultimately accelerate accurate drug research. Enterprise knowledge graphs excel at this situation, providing a semantic data layer to drive faster research. See Boehringer Ingelheim case study.
Another popular use case is to apply graph technology to predictive models of infectious diseases. When outbreaks occur, scientists can search various disease models, population data, and scientific research to quickly locate resources needed to halt the disease’s spread. See NIH case study.
Financial Services enterprises were also early adopters of graph technology. With graph technologies, financial institutions can build a single platform to comprehensively view firmwide parties, transactions, accounts, exposures, and various interrelated risks. This platform can power a range of use cases from regulatory compliance to Customer 360 to fraud detection.
Demands for data access are at an all-time high. Financial data is dynamic yet also deals with complex regulations. With graph, enterprises can:
- Adapt to changing regulatory and compliance demands
- Proactively manage IT assets to minimize risk
- Improve firmwide data management
Manufacturing & Automotive
Manufacturing is a complex web of product innovation, engineering, planning, production, and logistical functions. It is straightforward for organizations to lose value in the linkages between these elements due to data gaps, information latency, or even the need to make complex associations of data across the span of the operation.
With graph, manufacturers can instantly see the connections in their supply chain. They can enable scenario planning based on complete data, rapid identification of needed operational changes, and organizational resiliency. This drives long-term customer value and higher profit margins.
Unify data from product development, suppliers, manufacturing, and distributors regardless of the data’s structure, without moving or copying it.
- Represent complex dependencies in your supply network
- Reduce information blind spots
- Incorporate and act on valuable IoT data
- Digitize and optimize your supply chain
- Modernize Model-based systems engineering workflows
- Build the factory of the future
A fast-growing, newer cadre of startups is connecting data and delivering analytics solutions. They help solve urgent business problems, from optimizing retail shelf space to providing better patient care.
This wide variety of digital native organizations typically focus on data management and advanced analytics around a specific domain such as an industry and/or a competency across sectors such as strategic decision making or customer churn. They built their own data analytics platform, often comprised of a data lake and BI and data science capabilities. A growing number are employing knowledge graphs or a semantic layer to improve their advanced analytics capabilities.
For example, Caden uses Stardog’s unique semantic and data virtualization capabilities to unify consumers’ disconnected internet data. By applying multiple inference schemas and data models, they can create personal vaults that allow users to control and monetize their data.
And Capsifi uses a knowledge graph to help capture existing knowledge and connect it to a holistic model of the entire operation, revealing compelling insights that address the most challenging questions.
Energy & Utilities
Energy and Utility companies can also benefit from knowledge graphs. Monitoring and analyzing the topology of any given network of links and interfaces have become enormously complex, but these companies need to stay on top of it.
One use case is to create a near-real-time model of a utility’s network. This is what EXFO did. By linking network, service, and customer data within a graph, analysts can find the root cause of service outages, prioritize and deploy technical resources, and alert the affected customer within the same platform.
Other utilities may want to provide added value to customers through SaaS solutions. For example, Schneider Electric. Many sources of its building data (digitized floor plans, building management system, sensors, elevators, access control, etc.) link together using a knowledge graph, which maps the relationships between building spaces and a broad set of data points. A typical day results in more than 25 million events processed near-real-time, including events from hundreds of thousands of devices, sensors, and controllers.
Retail organizations were early adopters of graph technology. eBay is an excellent example of a digital innovator in the retail space. Its product listings (1.2 billion or so!) generate an enormous amount of data for analysis. eBay uses a knowledge graph to enable a continuous study of the most up-to-date data.
First-mile and final-mile delivery experiences also benefit from graph technologies. The FRONTdoor Collective, a unique network of partners specializing in these delivery experiences, has a flexible, reusable, and scalable data fabric layer. This layer brings together disparate data from its growing network of retail and delivery partners to improve routing efficiency, enable real-time shipment tracking, and mitigate operational risk.
Aerospace & Engineering
Engineering tasks are often complex endeavors of precision. They generally involve huge numbers of people, vast amounts of disparate data, and many interconnected pieces. And they leave no margin for error. Using graph technology, aerospace and engineering teams can quickly make informed decisions, improve engineering design, and reduce cost and risk.
NASA has many use cases involving graph. In this example, NASA faced a typical systems engineering hurdle; various divisions and contractors working across several disciplines required specialized data cuts from complex datasets. A knowledge graph swiftly provides the needed data to support mission-critical decisions.
Leveraging standards and specifications are also important to engineers. XSB connects engineers in the military and aerospace industry with the knowledge buried deep in the unstructured data of model specs and standards. By ingesting hundreds of documents weekly and connecting data concepts and lineage, XSB creates a “GPS for documents” that allows users to follow the trails of the spec they are tracing through time and space. This keeps the most up-to-date standard and spec versions in ‘engineers’’ hands and decreases the friction caused by static printed and pdf documents, reducing errors and risk across the enterprise.
No matter the use case, conventional, relational data models are hindered by initial assumptions about relationships between data — capturing only a fraction of all possible relevant relationships between data elements.
Graph greatly expands the captured relationships between data points, making it a superior data model for many use cases across industries. Graph projects can start with almost any dataset, including external data, making them the perfect choice to start small with and then scale up. Flexibility in data modeling and seamless schema extensibility provides the ability to accommodate newer data sources, data types, schemas, and use cases.