Pie Chart Percentage Incorrect

[AmberJane]

I can't post the data so will try my best to explain with screenshots. I have a dataset that is incident ticket, data, impact to client in mins and caused by. The caused by field is a multi-select divided by semi colons making filters and visuals with this field difficult. To get around this I duplicate the source dataset and split by delimiter on that field and link the two datasets. The issue with this is the data set that now has the nice clean caused by also has duplicate data. 

Left Table is the original Data Source and Right Table is the duplicate data but with the clean nice looking Caused By field: 

The issue I am having is I am using a pie chart that is using the caused by field from the duplicated datasource so it looks nice but the SUM of impact to the client from the original datasource. This looked correc to me as you can see from the screenshot below using application defect. The # of incidents and SUM of SLA being 4501 is correct: 

Now keep in mind I have multiple reports and datasets with multiselects that I set up this way. Built a beautiful dashboard and checked all of the fields and was ready to ship it. Until I checked the % calculated on the pie chart. That 27.31% you see for application defect is the 4501 minutes of impact divided by the total minutes of impact (using the duplicated data 😞 !!! ) not the original datasource :(. Which is a huge difference. 

Help Please!! 

[danextian]

Hi @AmberJane 
When you split the original category column by delimiter into multiple rows, you get distinct categories in each row. However, the amounts are also multiplied by the number of rows each category has been split into. To ensure that your pie chart displays the correct percentage, both the numerator and denominator need to be accurate. Unfortunately, altering the data label of a pie or donut chart isn't supported at this time. As an alternative, you could use a column visual that allows for a custom data label.

Please see attached sample pbix.

[AmberJane]

I tried following your pbix but couldn't understand how it compared to my data. A big issue I am having is I am using two datasets where your pbix only had one table. Thanks anyway! 

[danextian]

Because my dataset keeps the other original column so I can apply SUMMARIZE and SUMX on it to get the original value prior to expanding.

[danextian]

On a second thought, using the values like as if the categories were not split might be what you're looking for. Try using original value measure in the attached pbix.

19/75 =25.33%