Lesson 3 — The Six Modeling Schemes
Chapter 5 presents six major modeling schemes. The exam value is not deciding that one is “best”; it is recognizing the problem each scheme is designed to represent.
1. Relational modeling
Use relational thinking when the dominant concern is operational business rules, precise relationships, and stable data structure.
Recognition clues: - entities, relationships, attributes, keys; - operational applications; - normalization / “one fact in one place”; - crow’s-foot / Information Engineering style and names such as IE, IDEF1X, Barker, Chen.
A relational relationship line communicates business-rule structure, not merely a convenient analytic navigation path.
2. Dimensional modeling
Use dimensional thinking when the dominant need is analysis of a measured business process.
Recognition clues: - fact table; - dimensions; - grain; - conformed dimensions/facts; - slowly changing dimensions; - star or snowflake; - questions such as “Admissions Count by Semester, Campus, Program, and Region.”
Relational vs dimensional is one of the most important scheme distinctions:
Relational helps run the business accurately. Dimensional helps ask analytical questions about the business.
3. Object-Oriented modeling
Object-oriented / UML class modeling includes data plus behavior.
Recognition clues: - classes and objects; - attributes; - operations or methods; - encapsulation; - access visibility such as public/internal/private.
If the scenario emphasizes methods or behavior on a class, OO is the strong clue.
4. Fact-Based Modeling
Fact-Based approaches such as ORM/ORM2 and FCO-IM emphasize: - objects; - elementary facts; - roles; - controlled natural-language verbalization.
A key recognition clue is that Fact-Based Modeling does not use attributes as the primary modeling construct the way familiar entity-attribute models do.
5. Time-Based Modeling
Time-based approaches explicitly emphasize historization/evolution.
Chapter 5 examples include:
Data Vault
- Hub — primary business concept/key;
- Link — relationship/integration among concepts;
- Satellite — descriptive context and history.
Anchor Modeling
- anchors;
- attributes;
- ties;
- knots.
Do not swap the Data Vault vocabulary with the Anchor vocabulary.
6. NoSQL modeling
Chapter 5 treats NoSQL as storage-specific physical approaches. Recognition categories:
- Document — self-contained document structures.
- Key-value — retrieve value by known key.
- Column-oriented — organize large/sparse data by columns/families.
- Graph — nodes and connections; strong when traversal of highly connected relationships is the central requirement.
A scenario saying only “large data” does not automatically imply one NoSQL category. Match the shape of access and relationships.
Scheme-selection decision rule
Ask: What is the dominant representation problem?
| Need | Strong scheme clue |
|---|---|
| Operational rules and normalized structure | Relational |
| Measures analyzed by descriptive perspectives | Dimensional |
| Data plus methods/behavior | Object-Oriented |
| Facts/roles verbalized in controlled language | Fact-Based |
| Historized structural patterns | Time-Based |
| Document/key/value/column/graph physical shape | NoSQL |
Important exam trap
The six schemes are alternatives for different purposes, not maturity levels. One enterprise can legitimately use several.
Stop-and-check
- A system must enforce valid enrollment relationships during transactions. Strong first scheme?
- Executives need Enrollment Count by Semester and Campus. Strong first scheme?
- A requirement is “find nearest connections across an unpredictable number of people and devices.” Which NoSQL type?
- A model consists of controlled natural-language facts and roles and avoids attributes. Which scheme?
Answers: relational; dimensional; graph; Fact-Based.
Source anchors: current Mastery Lab scheme sections, pp. 135–143.