1. Home
  2. API Catalog
  3. Knowledge Graph Feed API (POC)

Knowledge Graph Feed API (POC)


Table of Contents

  1. What is Knowledge Graph?
  2. Why Knowledge Graphs
  3. What Content is Available
  4. Ontologies
  5. How to get the Knowledge Graph Feed
  6. How do I use the data?


Using verified Big Data management principles and machine learning techniques, the Refinitiv Knowledge Graph provides a trusted data source and model for customers to leverage, build their enterprise solutions on top of and benefit in new, connected ways from the breadth of the Refinitiv open platform.

By embedding the Knowledge Graph into company systems, you can deliver more powerful search and discovery applications to your users. The 2 billion relationships and data structures that power the Knowledge Graph enable financial firms to understand the complete picture of the ecosystem around their investments, targets, and prospects. Using the Knowledge Graph as a foundation allows customers to focus on their business’ opportunities and reduce their time spent on data management.

Used in conjunction with Refinitiv Intelligent Tagging, the Knowledge Graph serves as a foundation for the wider range Refinitiv Big Data solutions and is one of the core components of the enterprise content platform. Currently, the Knowledge Graph contains seven content sets including equity instruments and quotes, organizations, deals, supply chain, officers and directors, related companies and a comprehensive range of financial taxonomies and metadata. Over time, additional content will be added. Using linked-data principles of the Semantic Web, the Knowledge Graph publishes structured data in accordance with the Resource Descriptive Framework (RDF) and uses the open identifier standard Refinitiv Permanent Identifier (PermID). The Knowledge Graph feed is accessible via an API for delivery to customers’ on-premise or cloud infrastructures and is constantly refreshed as new entities and relationships are discovered.

With Refinitiv Knowledge Graph feed, you have the ability to start incorporating Refinitiv content as part of your bigger Institutional knowledge graph – connecting your data world to Refinitiv data and 3rd party data. Graphs can also be easily connected to other graphs as long as the graph databases share some common standards – typically around how entities (like people or companies) and relationships are represented.

Solve Business Challenges with Refinitiv Knowledge graph feed solution

  • Improve productivity of your Investment research function by enabling discovery of previously undetected relationships between entities, persons and events which have the potential to affect alpha.
  • Improve Business development and Customer Relationship Management (CRM) Systems through rich integration of internal, Refinitiv and third-party data enabling a graph network of sales leads and connections to be mapped out.


Every single one of us has access to the biggest most comprehensive set of information humankind has ever known. It's called the Internet. It is a phenomenon, an organically created snapshot of everything we could possibly want to know. But it is incredibly disorganized and chaotic.  

What if we could add a layer of meaning to the web, a lens so to speak that brings into focus the semantic meaning of all those different sources of information. And what if we could use that new common semantic understanding to make connections between pieces of data. To create a layer of overall knowledge from the connections and meanings.

That's what Linked Data represents - a common set of standards, tools, and practices to create sets of semantic meaning and to apply it to vast estates of data. To enable data to be described as collections of semantically organized statements rather than unstructured pages of text. And to enable collections of content to be weaved together through meaningful and commonly understood connections. To ultimately enable knowledge bases and to create more powerful ways to find, explore and discover the answers we really want to know.


Why is it important?

The internet is chaotic and segmented. It is a collection of silos and beyond the flexibility of HTML, there is no common language. The principal aim of Linked Data, in the outside world, is to provide a mechanism to make disparate sources of content interoperable. And to do so using a common standard approach, enabling tools that can run across and utilize great swathes of content.

This principle of interoperability creates a bedrock of linked content on which new solutions, products, and business models will be built. The next generation of online products will be based on knowledge engines and not just information.


What are we doing?

Linked Data is predicated on federated sources of data working together. It's important these sources of data, and the associated expertise within these domains, maintain ownership and control. So to 'play' in the Linked Data world just requires adherence to the common standards that describe it.

The Platform and Enterprise Architect groups, through the Linked Data initiative, will act as a hub to collate and represent a common set of industry standards and practices laying out how to describe and act on Linked Data. In this regard, it's important to ensure we properly understand the implications of adopting standards, best practices and ways to progress and improve standards, and this works best when consolidated.

However whilst federated development and ownership is good, it's also hugely important to decrease the cost of entry into the Linked Data world. In an area where we should encourage experimentation and the exploration of new functionality and user experiences, it makes sense to share and reuse as much as possible. To that end, the Platform group is creating a reference implementation and set of reusable 'components' to act as a starter kit.

The reference implementation also sets out to demonstrate a number of the key principles and opportunities in practice, for example: content can easily be transformed into Linked Data, ontologies can be defined to describe business knowledge, a variety of product facing databases (such as triple stores) can act on the same linked data and provide new functionality, and automated services can easily and consistently be applied across different domains of data.

And behind all this, we should continue to explore and innovate in this area, both technically and commercially. So a program of ongoing proofs-of-concept and prototypes will be supported to prove and demonstrate more of what is possible.


Refinitiv Knowledge Graph feed is a means of organizing and representing information. It comprises objects and relations between those objects, such that any pair of objects connected by relation form simple information. Our Knowledge Graph is based on the industry standard Resource Descriptive Framework (RDF) feed model. The RDF-based Knowledge Graph is already reaching billions of triples (Triple is: “john smith – is officer director of – company X”), containing attributes and relationships for entities – such as organizations, people, financial instruments and value chains.

Currently, the Knowledge Graph contains seven content sets including:

  • Equity Instruments & Quotes: 30 million equity instruments and 80 million quotes.
  • Organizations: 5 million Organizations, 1.1 million hierarchy relationships
  • Deals: Over 280,000 Investments and over 1,5 million Mergers and Acquisitions
  • Officers & Directors: Over 2.5 million people roles, biographies, career, education data
  • Comparable Companies: 60,000 companies with 1.5 million relationships.
  • Supply Chain:  35,000 companies with 55,000 relationships.
  • Meta-Data


Knowledge Graphs express all information as relationships. All Metadata is expressed as Identifiers which are shared between content types, allowing reliable connections to be made

Many data items in the graph are uniquely referenced by Refinitiv Permanent Identifier (PermID) which is our permanent and machine-readable identifier. PermID never changes, helping you handle complex data management challenges, eliminate mapping inconsistencies, reduce risk, and streamline end-to-end workflow processes across various platforms.


How Do Ontologies Work?

An ontology is a list of terms, properties, and classes that are being used. Furthermore, it’s a set of statement about those terms. Ontologies are written in RDF. 

Generally, an ontology will group items into a single namespace that describes a particular information domain - for example, there might be separate ontologies for People and Organizations. However, you can mix and match ontologies. Your data can (and will) include terms from many different ontologies.


What are the name spaces?

Before we can use a set of terms, we need a precise indication of what specific vocabularies are being used. A standard initial component of an ontology includes a set of XML namespace declarations enclosed in an opening rdf tag. These provide a means to unambiguously interpret identifiers and make the rest of the ontology presentation much more readable.


What is PermID?

Refinitiv Permanent Identifier (PermID) is a machine-readable identifier developed to create a unique reference for any data item. While most identifier schemas span subsets of entity types or categories, PermID provides comprehensive identification across a wide variety of entity types including organizations, instruments, funds, issuers, and people.

Unlike stock tickers and other such symbols, PermID never changes, helping you handle today’s complex data management challenges, eliminate mapping inconsistencies, reduce operational risk, and streamline end-to-end workflow processes across various platforms.


PermID Content and features

Identification capabilities

Provides comprehensive identification capabilities across a wide variety of information objects, including organizations, instruments, funds, issuers, and people.

Wide scope

Identifies a wide variety of object types, making it an ideal method for a better description of the relationships between those objects, and a descriptive anchor of an object’s properties or characteristics.


Acts as a machine-readable identifier, thus helping improve scale and reducing latency.

Connecting content

Connects Refinitiv best-in-class content offerings to generate insight, manage risk, and simplify various business processes.


How do I get it?

Request PermID Details here: https://www.refinitiv.com/en/products/permid-data-management/



  • The Graph Feed Administration interface is used to create and manage content
  • Any content residing in CM-Well can be extracted
    • Generally, queries are built around RDF types
    • We also use field-level parameters such as score thresholds
    • Finally, we apply SPARQL transformations, e.g. to remove system-level metadata or to add further linkages
  • ‘Reference’ content sets are built up from these underlying content queries and are available for delivery to customers now. We will continue to expand these
  • The Administration interface is also used to entitle consumers to content sets
  • The Graph Feed API is used by consumers to download and consume this content
  • An SDK is available to embed this functionality in customer workflows