Complete Power BI Data Modeling: Star Schema End to End

Learn Power BI data modeling from the first concept to a finished project: star schemas, relationships, fact tables, dimensions and row-level security. It is for Power BI users and data analysts who want a model that is fast and correct.

The course is 5.5 hours of video, 28 video lectures in 7 sections. The first 4 sections teach the concepts: first with hand-drawn sketches, then in Power BI Desktop on small tables. The last 3 sections are one project: you take a messy model and rebuild it into a clean star schema. It gets a date table, the core DAX measures and row-level security. Every step happens on your own computer, with the course files.

Most Power BI problems do not start in the report. They start in the model, and they show up as slow reports and totals that do not add up. The project starts with a messy model, the kind you find in a real company. You see what is wrong with it, and then we rebuild it from the raw data, the way dimensional modeling is done in large companies.

I have worked with data for 17 years, and I led one of the biggest data and AI projects at Mercedes-Benz. The steps in this course are the steps I used on those projects, not textbook theory.

What makes this course different:

• The concepts first, then one project from start to end: every lecture builds on the one before it

• Hand-drawn sketches explain the ideas before we click in Power BI Desktop

• The terms of Microsoft's documentation and of the job ads: fact table, dimension, grain, star schema, date table and row-level security

• My 2 models as .pbix files, one for the concepts and one for the project, so you can compare your work with them

What you build, section by section:

• Introduction: what Power BI is, how to install Power BI Desktop, and why data modeling matters

• Designing a star schema: star, snowflake and galaxy schemas, relationships, cardinality and filter direction, and your first star schema

• Relationships and dimensions: ambiguous relationships, one-to-one dimensions, junk dimensions and role-playing dimensions

• Fact tables and security: the grain of a fact table, several fact tables in one model, and static and dynamic row-level security

• The project, dimension tables: you diagnose the messy model, identify facts and dimensions, then build the customer and product dimensions

• The project, fact tables: you define the grain, build the sales, inventory and campaign fact tables, create the relationships and model the order fulfillment process

• The project, the finished model: the complete star schema, a date table, the core DAX measures, row-level security and a final validation

What you need: Power BI Desktop, which is free on Windows. A lecture shows the install, and the setup guide shows how to run Windows on a Mac with Parallels. No data modeling experience is needed.

What you get: the course files and my 2 models: 14 small data files for the concept lessons, 1 Excel file for the project, and a .pbix file for each part. The Excel file has 23 sheets: customers, products, orders of 2025 and 2026, invoices, payments, shipments, inventory, campaigns, exchange rates and sales targets. You also get the Q&A when you get stuck, and a certificate when you finish.

When you finish, you have a data model that you can show in an interview and reuse at work. If your reports are slow or your totals are wrong, this course shows you why, and how to fix it.

Enroll, download the course files in the Course Resources lecture, and build the model with me. See you in the first lecture.