This time we’re going bigger than ever. Fabric, Power BI, SQL, AI and more. We're covering it all. You won't want to miss it.
Learn moreDid you hear? There's a new SQL AI Developer certification (DP-800). Start preparing now and be one of the first to get certified. Register now
Hi,
I am looking for feedback on other experiences.
Today we build data models and reports in the same .pbix and deploy them. We are starting to use shared data source for multiple .pbix reports. I am comfortable with this.
In doing this, i was curious if it is a best practice to always have a backend data source and front end report file. Meaning, create 1 .pbix that is just the full prepped data model with all power quuery and dax completed, publish that, and then connect a 2nd file as the front end report file (even if you only plan to use the data source once).
Has anyone done this? What is your experience (e.g. are the refresh times impacted good/bad when developing, does it impact the user's speed experience, etc).
@Anonymous
It looks a good approach to me. But I also suggest that you also post this question in the "Power BI Service" forum
Check this article by @MattAllington
https://exceleratorbi.com.au/new-power-bi-reports-golden-dataset/
and this one
Meanwhile you can also explore "DATAFLOWS"
Here are recent articles by @reza_rad
http://radacad.com/what-are-the-use-cases-of-dataflow-for-you-in-power-bi
and
http://radacad.com/getting-started-with-dataflow-in-power-bi-part-2-of-dataflow-series
and a video by @ruthpozuelo
https://www.youtube.com/watch?v=bkFG8s_9sGE
Check out the April 2026 Power BI update to learn about new features.
Sign up to receive a private message when registration opens and key events begin.
If you have recently started exploring Fabric, we'd love to hear how it's going. Your feedback can help with product improvements.
| User | Count |
|---|---|
| 37 | |
| 32 | |
| 27 | |
| 24 | |
| 17 |
| User | Count |
|---|---|
| 70 | |
| 50 | |
| 31 | |
| 26 | |
| 24 |