What is denormalization from a modeling perspective?

< 1 min read

Denormalization means combining related tables into fewer tables to simplify reporting and improve query performance.
Instead of separate Customer, City, and Country tables, we merged them into one dimension table.
This reduced the number of joins needed in Power BI visuals.
Dashboard load time improved because aggregations ran on fewer relationships.
It also made the model easier for business users to understand.
There is some data duplication, but acceptable for analytics use cases.
We applied this while designing a star schema for sales reporting.
Source system remained normalized, but reporting layer was denormalized.
So it prioritizes performance and usability over storage efficiency.

Power BI Essentials