22-25 April 2026

Semantic Model Ideation for Power BI

Proposed session for SQLBits 2026

TL; DR

Building performant semantic models can be challenging with disparate sources. Learn my ideation process to create secure, best-practice models. We will work backwards from user questions to visuals to medallion architecture requirements for Synapse or Databricks.

Session Details

Building a performant star schema semantic model for Power BI enables simpler DAX, improved load time and responsiveness, and ensures that you are able to use the visuals that will best communicate your data. When starting with a snowflake, disparate data sources, or one big table, it can be a challenge to build something from the start that is going to fit your user requirements.

Enter the semantic model ideation process. I use this process to help clients create performant secure data models aligned to a star schema and Power BI best practices. Attendees will walk away with an understanding of what information they need to start this process, how to work through the steps, and how to validate the proposed model will support their visuals. We will work backwards from what data you need to support the user questions, what Power BI needs to support the visuals and front end performance, and an overview of the approach to building out your medallion architecture technical requirements for Synapse or Databricks to enable Power BI semantic models.

3 things you'll get out of this session

Attendees will learn how to think through planning their model, how this improves speed to delivery, and how it enables better feedback loops.