Module 3: Modelling Business Concepts

Estimated time: 75 min

This module is the core of the course. It is about modelling business concepts and their relationships instead of database tables: finding the concepts hiding inside a wide table, giving the important ones a class of their own, connecting them with relationships, and refining what auto-generation drafted.

Good modelling pays off twice. Users reach their answers with shorter, more natural questions, and the numbers behind those answers stay correct. Weak modelling does the opposite. Questions get harder to ask, and some results come back wrong.

What you will learn

• How to spot the business concepts hiding inside a single wide table.
• What classes, attributes and entities are, and which values are worth syncing.
• How relationships connect classes, with the right cardinality and direction.
• How an auto-generated Knowledge Graph is refined into a model users can trust.

Lectures

• 3.1 Think in Business Concepts, Not Tables - what one row represents, and why the grain decides the numbers.
• 3.2 Classes & Attributes - parent and attribute classes, the sql tag, and the attribute types.
• 3.3 Entities - why classes need entities, syncing, and the sync policies.
• 3.4 Properties - name_sql, synonym_sql and description_sql, plus the class-or-string test.
• 3.5 Relationships & Cardinality - connecting classes so cross-concept questions work.
• 3.6 Refining the Auto-Generated KG - from generated draft to a model users can trust.