Find everything you need to get certified on Fabric—skills challenges, live sessions, exam prep, role guidance, and more.
Get startedGrow your Fabric skills and prepare for the DP-600 certification exam by completing the latest Microsoft Fabric challenge.
Is there any good documentation on how to model my data to be consumed by PowerBI report?
I want to know what are some of the best practices that will allow me to explore data in a flexible way?
The Tabular engine underpinning the Power BI data model is largely model agnostic. Anything that can be represented well as a series of tables and relationships can function with the engine.
That being said, for ease of use and and consumption by end users, a dimensional model is far and away the best. I truly don't know of any better resources than the Kimball dimensional modelling books. If you're serious about building good models, you should read these. There are some things that I don't agree with the Kimball group on, but as a foundation, they can't be beaten.
Another worthwhile resource to check out is powerpivotpro.com, especially the archives, as some of the older posts would really dive into tactical design tips for the Tabular model. I don't find the site to be a good resource for general dimensional modelling, as that's not the goal. The goal is to advocate for Power Pivot and show value in specific business scenarios. If you can find an example that maps well to your own domain, then it is very helpful.
Another great blog worth reading is sqlbi.com. They cover some of the best modelling and design patterns specifically for DAX and Tabular.
While greggyb is correct, I think it is also a bit overkill to ask someone to read several books.
I would recommend you search online for a tutorial or introduction to "Star Schema". I found this, which at a glance seemed good. This is what greggyb refers to as "Dimensional modelling" and is the way to do it.
I think of them as an arrangment where facts are in the middle and dimensions around. Facts are tables of transactions, where nothing has to be unique. Dimensions are translation tables with background information where something has to be unique.