Why this is in my collection
From the publisher:
Get the best out of Power BI by building optimized data models for reporting and business needs
Highlights
- Report quality is decided at the model, not the visual — Bakhshi's whole premise is that an optimized star-schema model upstream determines everything downstream in Power BI; the MDDE claim (get the model right and the rest follows) applied to the BI layer.
- Transformation belongs in a defined layer — data preparation goes in Power Query or the source, analytical logic in DAX measures; assigning each kind of logic a designated home is layer discipline, the antidote to the untraceable everywhere-logic that makes BI solutions unmaintainable.
- Star schema as the load-bearing pattern — dimensional modeling is not one option among many but the shape the VertiPaq engine and DAX are built around; a reminder that pattern conformance, not cleverness, is what scales.
- Calculated column vs measure is an architectural decision — where a piece of logic lives changes storage, performance, and reuse; placement rules for logic are as much part of the model as the tables.
- Row-level security as declarative model metadata — access rules expressed inside the model rather than per report is rules-as-data: one definition, enforced everywhere it's consumed.
- Shared datasets and dataflows treat the model as a product — building the model once and letting many reports consume it is the reuse economics MDDE argues for at warehouse scale, played out inside the Microsoft stack.
Highlights on this page are generated with the help of AI.