Pistoia Alliance's FAIR & Data Governance Training: Technical Infrastructure - the FAIR Data Registry

Heiner Oberkampf, CEO and co-founder of ACCURIDS, introduced the essential role of a FAIR Data Registry in the Pistoia Alliance FAIR & Data Governance Training. He clarified various identifier challenges and emphasized that a shared FAIR data language with unique identification using URIs and GUPRIs is crucial for automation and AI.

‍

Defining GUPRIs

Heiner defined GUPRIs as long-lasting, clickable references governed by an authority. Citing AstraZeneca's holistic metaverse example, he detailed the implementation through data management systems registering equipment to create a GUPRI by combining an existing ID with a namespace.

‍

Building Blocks for Scale

To achieve FAIR data at scale, Heiner outlined building blocks starting with an enterprise-wide reference classification or taxonomy. He detailed three main steps to success:

‍

1. Registering things with semantic context.
2. Mapping existing data.
3. Interlinking data to build a scalable knowledge graph.

‍

He also outlined four different integration levels (L1-L4) between business applications and the FAIR Data Registry.

‍

Integration & Practical Adoption

Heiner concluded by comparing the FAIR data registry integration pattern to Gartner's MDM registry pattern for scalability. He illustrated practical adoption with AstraZeneca's production use, which currently involves:

‍

• Over 500 registered namespaces.
• More than 25 million GUPRIs under governance.