I asked Eric Tauer (adilas developer) to share with me a paragraph on some terminology used in data warehousing. Here is what I got from Eric on 11/16/19.

There are different technology approaches to Data Warehousing (DW).  My experience is related to the following...

A DW stores data for the purposes of on line analytical processing (OLAP).  The DW’s data sources are typically concerned with on line transactional processing (OLTP).  The physical data models used for OLAP and OLTP are structurally different.  OLTP is relational, at least to 3rd normal form.  OLAP is denormalized.  Denormalizing data has costs that benefit analysis of the data.  The cost of denormalized data is not beneficial for OLTP.  This means these two different data architectures are optimized for business analysis (OLAP) vs. transaction storage (OLTP).

Extract, transform, and load (ETL) processes are used to get source data from OLTP to OLAP.  During ETL, change detection, aggregation, denormalization, and more are exercised to store the data for efficient analytical uses.  A common physical data architecture for a DW is called a dimensional model.  Dimensional models are also called star schemas.  Dimensions (Customer, Vendor, Product, Time, etc.) are shared among many unique fact tables (Sales, Inventory, etc.).  Fact tables may also be a view of a logical group of tables that included multiple levels of aggregated granular fact data when and where performance gains are needed.  OLAP database schema design is optimized for the success for business intelligence, and involves a great deal of business analysis to make the many decisions in data modeling that will deliver a meaningful schema for data analysis.