Determine which column has the most commonality and what is it's percent of total

[datavigilante]

Hello all,

I am trying to analyze data on different outcomes and determine which factor is most common for a specific outcome.  To help explain, I've put some example data in the screenshot below:

I can use a slicer to show either the "Good" or "Bad" outcome records only.  I can also determine the most common value in each column using a technique I found in another topic, but I don't know how to compare the different columns to see which one has the most commonality (the green highlighted values).  I guess I would like the output to look like this:

Outcome	Most Important Factor	Percent of Total
Good	Factor 1 - C	50.00%
Bad	Factor 2 - E	83.33%

Thanks so much for your help!

David

[v-jiascu-msft · Accepted solution]

Hi David,

Firstly, you need to unpivot these there Factor columns into rows. Then you can try these measures below. There are two methods. You can check it out in this file. The following method would be easier.

Most Important Factor 2 =
CONCATENATEX (
    TOPN (
        1,
        SUMMARIZE (
            'Table1',
            Table1[Outcome],
            'Table1'[Attribute],
            'Table1'[Value],
            "Num", COUNT ( 'Table1'[Value] )
        ),
        [Num]
    ),
    [Attribute] & [Value]
)

Percent of Total 2 =
MAXX (
    SUMMARIZE (
        'Table1',
        Table1[Outcome],
        Table1[Attribute],
        'Table1'[Value],
        "percentagedd", COUNT ( Table1[Value] )
            / CALCULATE (
                COUNT ( Table1[Value] ),
                ALLEXCEPT ( 'Table1', 'Table1'[Outcome], 'Table1'[Attribute] )
            )
    ),
    [percentagedd]
)

Best Regards!

Dale

[datavigilante]

It will take me some time to absorb the concepts and then apply these techniques to my real data, but this looks great Dale.

Thanks so much for your help!

David