Master Power BI & Fabric

Microsoft Fabric is an end-to-end analytics platform that includes Power BI as one of its workloads. While Power BI focuses on reporting and visualization, Fabric adds data engineering (Data Factory, Lakehouse), data science (Notebooks), real-time analytics (KQL databases), and data warehousing (Synapse) — all sharing a single storage layer called OneLake. Think of Power BI as one room in the Fabric house.

The most impactful DAX optimization techniques include: (1) Use variables (VAR) instead of repeating calculations, (2) Avoid bi-directional relationships, (3) Use FILTER explicitly rather than passing complex conditions to CALCULATE, (4) Use IN {} instead of nested OR operators, (5) Minimize ALLSELECTED usage, (6) Always use a proper date table with marked relationships, and (7) Move calculated columns to Power Query where possible.

The Medallion Architecture organizes data into three layers: Bronze (raw data ingested as-is, typically as Parquet files), Silver (cleansed, validated, and conformed data stored as Delta Lake tables), and Gold (business-level aggregations and star-schema tables optimized for Power BI consumption).

OneLake is the unified SaaS data lake at the core of Microsoft Fabric. All Fabric workloads read and write to OneLake without data movement. OneLake Shortcuts let you create virtual links to data in Azure ADLS Gen2, Amazon S3, Google Cloud Storage, or other Fabric workspaces — and query it as if it were natively in your Lakehouse, without copying.

If the transformation doesn't depend on the visual's filter context, do it in Power Query. Power Query runs once at refresh time, reduces model size, and keeps DAX measures clean. Use DAX only when you need calculations that respond to slicers, cross-filters, or other user interactions.