Star Schema Modeling

DW implementation in SIGADW usually occurs as follow:

Affected parties define what they want to analyze, always according to management analyses and never according to operational analyses. Interviews with affected parties are guided by IT area of the organization or by external consultants, that analyze the importance of indicators ordered.

Once made the whole logical survey of needs, next step is to design how DW is, through the modeling known as Star Schema. In this modeling, what will be analyzed is defined, that is, breaks and indicators. Breaks, called dimensions, define the entities to be analyzed (for example: products, customers, dates, etc.) and indicators are the numbers.

The schema can have related information about different areas, such as financial, invoicing, accounting, etc. It is the essence of a DW. If the model has data of only one company sector, it is called datamart, that is a subdivision of data warehouse.

Once mapped Star Schema model, just define where data of each dimension and indicator are. Then, you must run extraction, transformation, and load of data.

These data can be spread in several systems in an organization. They are also called data sources, read from their systems and stored in a big database, called data warehouse.

It is not the operational database, but it is a database that is used from time to time. These operations are made by IT area of the organization, that has all knowledge of systems and data that are loaded for SIGADW.

These data extracted, transformed, and loaded for data warehouse are then used to generate management queries.

These queries can be created by either SIGADW administrator or final users.