Dimensional Data Modeling - Semantic Layer

About

A semantic layer is a semantic data model that shows to the business user:

• a dimensional view (dimensional schema) from
• a relational schema.

It will create/model a logical star schema by:

• tagging:
  • the tables as dimension or fact table
  • the columns as measure or description
• creating derived measure (calculation)
• defining:
  • the relationship between tables
  • the dimensional hierarchy and its level

Once the semantic layer has been build, the business user will see a logical star schema that it can query (business query). This query is called a semantic query.

The semantic layer will translate:

• a semantic query
• to a sql query.

It's a key components of business intelligence / data analysis application because it provides a translation of the underlying database structures into business user oriented terms (such as revenue by year, …)

Articles Related

Example

• the logical business model for OBIEE,
• the universes for Business Object

Data Modeling Issue

While modeling the semantic layer, the developer can fall in different issues. See Semantic Layer Modeling - Issues (SQL Trap)

Documentation / Reference

• Design Tip #158 Making Sense of the Semantic Layer