Ben Richardson explains a common anti-pattern in Power BI semantic models:
Relationships sit at the heart of literally everything you do in Power BI.
Before you make measures, visuals and reports, relationships are established to define how your data fits together. Their job is simple on the surface – but vital: describe how each table is connected.
If you can design these relationships well, everything else will run much smoother.
Across any data domain, strong models rely on clear Grain, correct Cardinality, and a Star Schema built with well-defined Fact and Dimension tables.
Read on to understand how many-to-many relationships stress this understanding in Power BI an different techniques for dealing with those sorts of relationships.