Share

Dataflows and Datasets in Power BI

Christopher Wagner • Jan 04, 2020

It's Time to get your FLOW on

Dataflows and Datasets are some great new features that you want to leverage moving forward. They are mighty and enable several great features that add resiliency, ease of access, lineage, auto-documentation, and several other benefits that enhances ROI to each business unit.


What are Power BI Dataflows? Power BI Dataflows allow you to define individual tables that can be used in different data models out in Power BI. Some examples would be a Product, Employee, Date, or Transactions table that you would want to use the same information in different data models. Dataflows allow you to load the data from the source system out into the Power BI service a single time, transform the data once, and then consume the data many times.

Best Practices

Create a Power BI Dataflow


Dr Dataflow himself, Matthew Roche explain the difference between Power BI Dataflowsand Power Platform Dataflowsin this great video that he has over at BI Polar.

What are Datasets? Datasets are a combination of tables, joins, and measures that can be used to build out Power BI reports. Any time you build out a Power BI report, you are building a dataset. Radacad has a great article on what is a Dataset and how can you use them to improve your reporting and performance.

https://radacad.com/power-bi-shared-datasets-what-is-it-how-does-it-work-and-why-should-you-care


Why should you use a common Dataset? Excelerator BI published an excellent article over a year ago about the value of having a "Golden Dataset" with an easy to understand examples. Check out the first two sections in this blog ("The Problem- Too Many Data Models" and "What is the Golden Dataset"). The rest of the blog provides a detailed technical guide that is nearly 100% out of date and should be ignored.

https://exceleratorbi.com.au/new-power-bi-reports-golden-dataset/


What are Promoted & Certified Datasets? Once you start to build out datasets that can be used for multiple reports, you can category the dataset as either "Promoted" or "Certified". Best Practice is to mark a dataset that are used by a team or division as "Promoted" and any dataset used by the enterprise as "Certified".

Our friends over at Guy in a Cubehave a great video about Shared & Certified datasets.


CHRIS WAGNER, MBA MVP

Analytics Architect, Mentor, Leader, and Visionary

Chris has been working in the Data and Analytics space for nearly 20 years. Chris has dedicated his professional career to making data and information accessible to the masses. A significant component in making data available is continually learning new things and teaching others from these experiences. To help people keep up with this ever-changing landscape, Chris frequently posts on LinkedIn and to this blog.
By Christopher Wagner 15 Nov, 2023
In a dynamic data engineering scenario, Sam, a skilled professional, adeptly navigates urgent requests using Microsoft Fabric. Collaborating with Data Steward Lisa and leveraging OneLake, Sam streamlines data processes, creating a powerful collaboration between engineering and stewardship. With precision in Azure Data Factory and collaboration with a Data Scientist, Sam crafts a robust schema, leading to a visually appealing Power BI report.
By Christopher Wagner 28 Apr, 2023
NOTE: This is the first draft of this document that was assembled yesterday as a solo effort. If you would like to contribute or have any suggestions, check out my first public GIT repository - KratosDataGod/LakehouseToPowerBI: Architectural design for incorporating a Data Lakehouse architecture with an Enterprise Power BI Deployment (github.com) This article is NOT published, reviewed, or approved by ANYONE at Microsoft. This content is my own and is what I recommend for architecture and build patterns.
Work Hard - Let's GO!
By Christopher Wagner 14 Apr, 2023
Work Hard - Let's GO!
Data God Guide to Learning DAX
By Christopher Wagner 22 Jan, 2023
DAX is more than the name of my new puppy
Show More
Share by: