Glossary/Semantic Layer & Metrics

Universal Semantic Graph

A universal semantic graph is a unified representation of an organization's data entities, relationships, and metrics that serves as a single reference point for analytics across all tools and use cases.

A universal semantic graph (sometimes called a knowledge graph in analytics contexts) maps all business entities (customers, products, orders, accounts, transactions) as nodes and their relationships as edges. It includes attributes, hierarchies, and calculated metrics. Rather than maintaining separate data dictionaries, semantic models, or metadata repositories, a universal semantic graph consolidates all semantic information into one graph structure that analytics tools can query and reference.

The universal semantic graph solves enterprise fragmentation: different business units define "customer" differently, line-of-business teams maintain separate BI environments, and metric definitions diverge across regions. A universal semantic graph enforces one canonical view. It enables discovery: analysts can navigate from a customer to all related orders, products, segments, and metrics via the graph. It also enables cross-domain analytics: tracking how decisions in one domain (e.g., marketing spend) affect another (e.g., product adoption).

Universal semantic graphs are typically implemented as graph databases, knowledge graphs built on data warehouses, or catalog systems with rich relationship tracking. They integrate metadata from multiple sources (tables, dashboards, queries, documentation) into one queryable structure. The graph is living documentation: staying current as the business evolves rather than becoming stale.

Key Characteristics

  • Represents entities as nodes and relationships as edges
  • Includes business metrics, dimensions, and hierarchies
  • Queryable and discoverable across the organization
  • Integrates metadata from multiple sources
  • Tracks lineage between entities and metrics
  • Supports role-based access and governance

Why It Matters

  • Discovery: Find related metrics, entities, and lineage instantly
  • Consistency: Single definition of core entities prevents fragmentation
  • Trust: Transparent relationship mapping builds confidence in analytics
  • Scale: Support thousands of entities and metrics without redundancy
  • Governance: Track who owns what and enforce policies at scale

Example

A universal semantic graph connects: customer entities to orders, orders to products and shipments, products to categories and suppliers, and metrics (revenue, churn, lifetime value) to their underlying dimensions and calculation rules. An analyst exploring churn immediately sees which customers, products, regions, and cohorts are affected without navigating separate systems.

Coginiti Perspective

SMDL's entity-relationship model forms a semantic graph where entities are nodes, relationships define edges with explicit cardinality, and dimensions and measures annotate each node with business meaning. Semantic SQL traverses this graph to resolve implicit joins, so a query referencing attributes from multiple entities navigates the relationship paths automatically. Because SMDL files connect to 24+ data platforms, the semantic graph can span entities across different databases and cloud providers within a single queryable model.

See Semantic Intelligence in Action

Coginiti operationalizes business meaning across your entire data estate.