Reference Data Types and Their AI Value

Different types of open source reference data serve different enrichment purposes:

Occupational taxonomies (O*NET, ESCO, SOC)

• Enable AI to understand job roles, required skills, task structures
• Support HR, workforce planning, talent management assistance
• Bridge between informal job descriptions and standardized categories

Knowledge graphs (Wikidata, DBpedia)

• Provide entity metadata for disambiguation
• Help AI distinguish between “Apple” the company and “apple” the fruit
• Rich relationship data for contextual understanding

Skills taxonomies (ESCO, OSMT)

• Map competencies to occupational requirements
• Enable skills-based matching and gap analysis
• Support career development and workforce planning

Technical ontologies (Common Core, CodeOntology)

• Structure knowledge about software, engineering, systems
• Enable more accurate assistance with technical documentation
• Provide domain-specific vocabulary and relationships

Commonsense knowledge (ATOMIC, ConceptNet)

• Capture everyday reasoning patterns
• Help AI understand causes, effects, social dynamics
• Fill gaps that technical knowledge doesn’t cover

The pattern: each type of reference data enables a different class of AI capability. The combination creates compound enrichment.

Related: 04-molecule—reference-data-multiplier, 06-atom—entity-linking-dimensionality, 06-molecule—knowledge-graph-construction