Diverging Stacked Chart from One Column

Question

I am attempting to create a diverging stacked chart based on a 4-point likert scale (SD, D, A, SA). I can easily create a stacked bar chart, as shown in the first screenshot below.

However, I would like to create a diverging stacked bar chart, as shown in the 2nd screenshot, which is an example from another visualization that has nothing to do with my data. In my example, there would be one single line dividing the red from the blue.

Here is the problem. All of the examples I read are based on each scale having its own column (four colums). One column for agree, one for strongly agree, and so forth. However, I pivot my survey data into one column and then filter the data using tables.

Here is what I have now. The 1st screenshot shows my model view. I use the demo table to filter responses based on demographics. The 2nd screenshot shows my data table view, showing the pivoted data table. I also have a sample, de-identified sample PBIX and Excel file here: https://drive.google.com/drive/folders/1Sbr-Nmnvou5J-k9KNEG-X8_KTA5K_mJQ?usp=sharing

I know I most likely to create a measure that multiplies the values by a negative, but I can't figure the DAX. This community is awesome and I would appreciate anyone who has a solution to this - thank you so much!!!

rlsmith1994 · Answer

Thank you for the response! I did&nbsp; consider the clustered bar chart. However, I would prefer the diverging stacked chart because it displays positive and negative sentiments. SD/Disagree are negative sentiments, SA/Agree are positive sentiments. In the clustered bar chart, all sentiments are above the zero line and thus "positive."&nbsp;&nbsp;I will play around with it a little more and may go with the clustered bar chart in the end. Thank you for your help!!

anonymous · Answer

Hi&nbsp; rlsmith1994&nbsp;,
&nbsp;
Are you referring to representing the table Agreement[Response] with four columns in each Status[Response] grouping?
You can consider using Clustered bar chart .In visual filter - Data lables - open Value.But I don't know what is your logic for %GT Count of value in the original visual because when I was looking at the data, I realized that Agreement[Response] and Status[Response] are equivalent to the same level.Can you describe in detail the Calculation logic? We can help you better.