Dimensional Modelling

Product Description

Overview

Dimensional modelling is an integral part of any BI (Business Intelligence) system & can be used within the data warehouse and/or the data marts.
This 3 day class assumes no prior knowledge of dimensional modelling. It starts by discussing what a data warehouse is, how they are designed & the part that dimensional modelling plays.
The vitally important process of requirement gathering is covered & attendees are shown how to:
Collect the analytical requirements of the business users
Create a logical model of these requirements
Create a star schema from those requirements
The relational & dimensional models are compared & contrasted, with particular reference to the current Kimball/ Inmon debate.
The class then looks in great detail at dimensional modelling itself & finally ends with a summary of possible BI architectures.
Audience
This class is aimed at people who work in the BI area. It is suitable for business analysts who need to understand the analytical requirements & turn those requirements into a model. It is also suitable for the IT professional who will turn those models into working OLAP (On-Line Analytical Processing) structures.
Class outline
Introduction to designing dimensional data warehouses
Gathering analytical requirements
Logical (Sun) modelling
Physical modelling ? the star schema
Facts & dimensions
Attributes & hierarchies
Time dimensions
Synonym dimensions
Surrogate keys
Additive, semi-additive & non-additive measures
Degenerate dimensions
Slowly changing dimensions
Bridge tables
Mini dimensions
Hot-swappable dimensions
Multi-valued dimensions
Parent child dimensions
Bitmap dimensions
Ragged hierarchies
Unbalanced hierarchies
Step dimensions
First & last analysis
Optimizing fact table performance
Indexing in star schema
Aggregation
MOLAP
HOLAP
ROLAP