How to identify unused objects in Power BI

BI, Tools, UMT360

This is a combined solution that you can use to identify unused objects in a pbix file. The scope is to identify those measures and columns from your report that are not used at all and could possibly be removed.

Required tools:

  • Power BI Desktop (the report I need to analyze);  
  • DaxStudio;
  • Performance Analyser (Power BI);
  • A final Power BI report to combine all the data.

Steps we need to follow to gather the necessary information:

Here I can run different queries, but I didn’t find this enough for what I need, so I will only take some results and use them in the final solution.

  • Get the .json file from Performance Analyser (Power BI)
    • We need one empty page in our report;
    • Go to View-> Performance Analyzer and make the pane visible;
    • Click Start recording;
    • Go on each page and wait until all visuals are fully loaded;
    • Press Stop Recording
    • Export JSON file
  • Open our custom NotUsedObjects report (do not close DaxStodio or our PowerBI report to analyze)
  • Here there are two parameters to change:
    • The process ID of the Analysis Service instance (it’s the one from the DaxStudio at the bottom of the page)
    • JSON path/name
  • Then we need to apply the changes:
  • And click on Run in the following dialogs:
  • We will use “Windows” and “Use my current credentials”:

In our final report, I have created 3 pages: for Measures, for Calculated columns, and another one combining both measures and calculated columns.

Please note that this method does not consider the measures and calculated columns used in Query Editor or Filter Pane, so you need to take a look in both places before you decide to remove an object.

This method allows us to combine useful outputs from DaxStudio, Power BI Performance Analyzer and Power BI and have it all in the same place. It offers a clear view of what objects we are not using anymore in our reports and helps us clean up our data model.

Click here to download the sample solution.

Cat image credit

What do you think?