Choosing a Data Model Diagram

Classify First

Tree of one-to-few, loaded as a whole: document model
Many-to-many with multi-hop queries: graph model
Mixed relations, joins, and analytics: relational or star schema

Relational Strength

Joins and many-to-one or many-to-many relationships
Schema-on-write enforces structure on every row
Star and snowflake schemas dominate analytics warehouses

Document Strength

Self-contained trees with read locality
Schema-on-read suits heterogeneous data
Limits: no direct reference to nested items, weak joins

Trade-offs and Convergence

Normalized: cheaper writes, joins on read
Denormalized: fast reads, expensive consistent writes
Modern databases blend models: JSON in SQL, joins in documents