Intended Audience: Confident Beginners, aspiring Intermediates
We will explore the underlying data model created by the enhanced Northwind trader data set. We will review the fields, tables, relationships, and measures that have been built into the model. The class will cover, in detail, facts and dimensions, data model schemas, cardinality and grain.
- Explore the data model as it is loaded into the Desktop
- Examine the relationships that Power BI created by default, and which ones are not present
- Return to the Power Query Editor to modify some data, and reduce cardinality of some fields
- Discuss how creating staging queries can ‘slim down’ a data model
- Hide fields in the data model to ensure users can’t misapply them
- Review relationship directions
- discuss how DAX filters traverse the model
Next, we will build a simple report and add some simple DAX measures and then measure the performance of the report. We will then perform some ‘tuning’ and measure the report again. The focus of this section will be interactive discussion of the pros and cons of the tuning techniques.
- Identify ways to improve the DAX measures
- Setting slicers with defaults
- Limiting interactions within visuals and across pages
- Avoiding high levels of granularity
- Using tooltips judiciously
- Using filters rather than slicers
- Testing Row Level Security roles
Conclusion
- Principles for good report performance
- Good resources for further learning
General Notes:
- This is a four-hour session designed to be delivered in two segments of two hours each.
- This session will not cover DAX performance tuning in great detail.
- Participants will need to have the Power BI Desktop installed.
- Data files (including a PBIX) will be provided prior to training.