Optimizing a Power BI Data Model

In Optimizing a Power BI Data Model we start by investigating the underlying structure and operation of Power BI's VertiPaq database engine. Armed with this knowledge, we investigate various strategies for optimizing performance. Not to give the game away, most of the strategies involve reducing the size of the data model and loading pre-summarized data. We also look at tools within Power BI and Power Query Editor for analyzing report and query performance as well as internal error tracking and diagnosis.

Learning Objectives

• Remove unnecessary rows and tailor data sets for report audiences
• Change data types to improve storage and split columns into multiple columns to reduce the overall number of unique values
• Remove redundant or duplicate data and replace columns with measures where possible
• Use Performance Analyzer to highlight report bottlenecks
• Create aggregations for imported and direct query data models
• Configure and use Power BI's and Power Query's diagnostic tools

Intended Audience

This lesson is intended for anyone who wants to improve the speed and responsiveness of their reports by improving the underlying data model.

Prerequisites

We would highly recommend taking the Developing a Power BI Data Model and Using DAX to Build Measures in Power BI lessons first, as they go into depth about how to implement several of the topics discussed in this lesson.

Resources

You can find datasets used in this lesson in this GitHub repo cloudacademy/get-data-into-powerbi. Consult the readme to down load the appropriate file(s)