top of page

Measure Killer Analytics
(preview)

part of external tool for Power BI "Measure Killer"

Calculations

​

The first feature of the new analytics is the overview of calculations (measures and calculated columns). It can show you if a calculation is used, the number of uses, the number of artifacts used by the calculation, and the estimated size of the model it uses. Also, this tool will show all the hierarchical relationships between the different artifacts in your model, making it easier to map out the artifact usage.

​

The table has 7 columns: 

1- Used

2- # of uses

3- Art. used

4- size (MB)
5- distinct (MB)

6- % of the model

7- % distinct

​

The used column tells you if a calculation was used, and # of uses is the number of times the calculation was used.

Art. used is the number of artifacts (other calculations or columns) that are directly or indirectly used for this calculation.

The size (MB) is the size of all the referenced items of the calculation added up. If a calculation uses another calculation that uses a column, this also counts towards the size (MB). This is because MK is looping through all levels of use to find all artifacts that contribute to a calculation. 

The distinct (MB) is the size of all the referenced items that were only used for one specific calculation and nothing more. It is the size that could be freed up by removing the calculation and its references. 

For % of the model and % distinct, the MB of the corresponding values (size (MB) and distinct (MB)) are divided by the total MB of the model giving the percentage of the total model. 

 

In the top left corner is a search bar where you can look for specific calculations. If you only want to see the unused calculations, you can click the Filter unused button in the top right corner. If you want to see all calculations again click the Reset table button and to delete everything you can press Clear Table. The ? button will guide you to the documentation. 

​

Definition of distinct:
Calculations that reference a column that are only used by that calculation and nothing else.
That means that removing the calculation would make it possible to remove the column without issues.
Removing the calculation could still break other calculations that references it.

​

Reports/pages/items:

​

The second feature of the new analytics is “reports, pages and items”. Here, you can map the columns and calculated columns of your model to their “final uses”, basically showing the relationship between the highest and the lowest level of the usage hierarchy, removing the middle artifacts.

 

“Final uses” are defined as something that is used by the report itself, and not by another artifact: active or inactive relationships, visuals, calculated tables, slicers, M expressions or column sorting. Calculated columns and measures aren’t considered “final uses”, because they should eventually be used by something else in the report. This means that if a column is only used in a measure or calculated column that is not used in any of the cases listed before, the column won’t show up in the analysis, since it’s not contributing to the report. 

 

There are three “distinct modes” to choose from report, page, and item. This can be done with the Distinct mode dropdown. The distinct modes change what is considered distinct.

​

-> In the report mode, every column or calculated column that has final uses only in one report is considered distinct. This means it can have multiple final uses as long as they are all within the same report. Removing the report would make the columns listed as distinct removable from the model. 

-> In the page mode, every column or calculated column that has final uses only in one page is considered distinct. This means it can have multiple final uses as long as they are all on the same page. Removing the page would make the columns listed as distinct removable from the model. 

-> In the item mode, every column or calculated column that has final uses only in one item (relationship, visual, calculated table, slicer, M expression and column sorting) is considered distinct. This means it can have multiple final uses as long as they are all within the same item. Removing the item would make the columns listed as distinct removable from the model.

 

If you delete something distinct, there is still the possibility of a measure or calculated column breaking but only if it is not used in a “final use” case.

​

To visualize all this, you have a table that is split up into 4 levels, “Report”, “Page”, “Item” (relationship, visual, calculated table, slicer, M expression and column sorting), and “Artifact” (Column or Calculated Column). All rows that include distinct data are colored blue. If no distinct data is included, the rows in the first three levels are white, and the rows on the artifact level (column or calculated column) are grey. 

​

The table has 5 columns:

1- Type

2- size (MB)

3- distinct (MB)

4- % of the model

5- % distinct  

​

The type column shows you the level the rows is in. For the first two levels (report, page) it gives you the name of the level and for the other two levels (item, artifact) it shows you which kind of item (relationship, kind of visual, etc.) or artifact (column, calculated column) is used.

The size (MB) is the size of the artifact. The distinct (MB) is again the size of the artifact but only if it is distinct.

For % of the model and % distinct, the MB of the corresponding values (size (MB) and distinct (MB)) are divided by the total MB of the model giving the percentage of the total model. 

 

This enables the user to see the cost (in terms of storage) of a report, page, or item, meaning if you delete a certain page or item, you save an X amount of storage.

​

Buttons

​

In the top left corner is a search bar combined with a dropdown where you can look for specific things. First you have to choose what you are looking for, report, page, visual or artifact. Then you just type in the characters you are looking for and press “Enter”.

If you only want to see the distinct rows, you can click the Filter Distinct button in the top right corner.

If you want to see all rows again click the Reset table button and to delete everything you can press Clear Table.

With the Expand/Collapse button you can either expand to open up all levels or collapse to only show the reports on hand.

The ? button will guide you to the documentation. 

​

bottom of page