Linked Data

Every distinct entity—whether a person, place, concept, or physical object—must be identified by a unique, dereferenceable Uniform Resource Identifier (URI). This replaces ambiguous strings like 'John Smith' with a globally unique identifier such as http://dbpedia.org/resource/John_Smith. Using HTTP URIs ensures that the name is not just an identifier but a locator on the web, enabling clients to look up the definition of the thing itself. This principle is the foundational step in moving from a web of documents to a web of things.

The identifiers assigned to entities must be HTTP URIs so that people and agents can look them up. When a client dereferences the URI, the server should return a representation of the identified resource using standard web protocols. This is the mechanism that transforms a simple name into a gateway for discovery. Without this principle, URIs remain opaque identifiers; with it, they become active links in a global information network, enabling the retrieval of structured data via content negotiation.

When a URI is dereferenced, the server must return useful, structured data about the resource using a standard format. The primary standard for this is the Resource Description Framework (RDF), which models information as triples (subject-predicate-object). Other serialization formats include:

• RDF/XML: The original XML-based syntax.
• Turtle (Terse RDF Triple Language): A human-readable, compact syntax.
• JSON-LD: A JSON-based serialization ideal for web developers. This ensures that the data is not just human-readable HTML but machine-processable information.

The final and most powerful principle is to link the data to other resources by using their URIs. By connecting a person to their place of birth (http://schema.org/birthPlace) or an organization to its headquarters, data is woven into a Web of Data. These typed links allow agents to discover new relationships and navigate across disparate datasets seamlessly. This transforms isolated data silos into a single, global database where the value of the data increases exponentially with each new connection, enabling complex federated queries.

Global manufacturers use Linked Data platforms to track components across multi-tier supply chains. Each part receives a unique URI, enabling provenance queries that span organizational boundaries without centralized databases.

• EPCIS standard uses RDF to model event-based supply chain data
• Enables queries like 'Which batches contain this defective component?'
• Supports regulatory compliance for pharmaceuticals and food safety

Europeana, the European digital library, aggregates metadata from thousands of museums, archives, and galleries using the Europeana Data Model (EDM), an RDF-based ontology. This interlinking reveals connections between artworks, artists, and historical periods that would remain hidden in siloed catalogs.

• Links a painting in the Louvre to a manuscript in the British Library via shared creator URIs
• Uses SKOS for multilingual thesauri and controlled vocabularies
• Enables serendipitous discovery across institutional boundaries

The Linked Open Drug Data (LODD) project connects disparate pharmaceutical datasets—from clinical trials to gene expressions—using RDF. Researchers can traverse a single graph to answer complex questions about drug interactions, target proteins, and side effects.

• Integrates DrugBank, PubMed, and ClinicalTrials.gov
• SPARQL queries replace manual cross-referencing across databases
• Accelerates hypothesis generation for drug repurposing

Governments worldwide publish statistical data, legislation, and geospatial information as Linked Open Data. The UK's data.gov.uk and the EU's data.europa.eu use RDF to make public sector information machine-readable and interlinkable.

• Links budget line items to geographic regions via shared URIs
• Uses DCAT vocabulary for cataloging datasets
• Enables journalists and civic hackers to build accountability tools