Semantic modeling glossary

Term

Definition

Data Hub vs Industry Term

Link

Dimension

Dimensions are tables representing business entities, for example Customer. Dimensions support filtering and grouping.

Industry

Attribute

Attribute

Attributes are the descriptive columns of a Dimension.  Day of Month in a Date dimension is such an example

Industry

Attribute

Calculated Member

A reusable formula whose value is calculated using other member values (for example, by using a mathematical or logical operation). The member is dynamically calculated at the time of use.

Data Hub

Calculated Members

Cube

A multidimensional view of the data.  It is presented in SSAS and is only relevant to the dimensional semantic layer.

Industry

Cube

Measure

Measures are the observation or event columns of a Fact table. Measure columns are almost always numeric.

Industry

Measure

Fact

Facts are tables that store observations or events, for example transaction amounts associated with an order. Facts support summarization.

Industry

Measure

Member

A single data value or cell in an OLAP cube, occurring at the intersection of data levels, attributes, hierarchies, and dimensions.

Industry

Member

OLAP

Online Analytical Processing

Industry

OLAP

Star Schema Design

The "Star" schema refers to design where one Fact table can relate to many Dimension tables. This forms a "star" concept. For example Sales Orders Fact can relate to Dimensions: Customer, Product, Account and Date. The star schema consists of one or more fact tables referencing any number of dimension tables.

Industry

Star Schema

Semantic Layer

A Semantic Layer is a modeling layer where several approaches for data analytics is available.  In Data Hub two approaches exist currently.  A dimensional semantic layer for use with prebuilt Data Hub analytics, and a tabular semantic layer for use by PowerBI and other 3rd party tools

Data Hub

The complete guide to the Semantic Layer

Semantic Model

The Semantic model in Data Hub can refer to either the "cube" when talking about dimensional modeling, or the tabular model.  It is the object that stores the physical data in its optimized reporting  form.

Data Hub