Effective model design
The concepts of relationships and filter propagation enable powerful analytics in Power BI models. It is important to think about the design of a model: which tables should the model contain, and which columns need to be in those tables? Which relationships are needed? In short, what is the overall structure of the model? The choices you make in model design will determine what results the model will be able to deliver.
Star schemas and snowflakes
A best practice for analytics using relational databases is to work with a specific database structure, known as a star schema. The basic ideas of star schemas apply to Power BI models as well.
Figure 1.2.12: Generic star schema structure
The central tables in a star schema model are the fact tables. These tables contain things that have happened, will happen, or should happen; like sales transactions, financial ledger transactions, customer inquiries, student enrollments, sales opportunities...