Data Modeling Best Practices

This article briefly provides the best practices for data modeling.

Detailed Information

Expert modelers always follow these rules: 

• Avoid wide tables. Smaller is better.
• Use a star schema. Always.
• Do not use bidirectional relationship unless you have a strong reason to use it.
• Hide all the columns in the fact table. Only show measures in the fact table.
• Expose visible attributes only through dimensions.
• Test your formulas on small sets of data that you can master and understand at a glance.

Expert modelers may decide to override the rules, but they do so with a deep and thorough understanding of what they are dealing with. Inexperienced modelers oftentimes choose to avoid rules. Doing so is exciting, if you want to feel the thrill of wandering into a totally unexplored and dangerous maze of complexity. On the other hand, if you want to deploy a sound model for your customers, then obeying these simple rules is a very good step in the right direction.

Resources

• For even more best practice rules, check out this article, updated in May of 2021, on the Microsoft Power BI Blog.

• From Microsoft: Understand star schema and the importance for Power BI

• Guy in a Cube video
  
  
  

• sqlbi free modeling course

  • Introduction to Data Modeling for Power BI

  • Understanding Relationships in Power BI

• Golden rules, source 
  • The Importance of Star Schemas in Power BI