Re-scaling a parameterized measure for each row in data

[Anonymous]

Hi all,

I built a simple index for each row in my data set that is a weighted linear combination of several columns. The weights are what-if parameters that the user can input and adjust to see the index change on the fly. Now, I want to rescale this raw index into a score out of 100. Conceptually, the mathematics for that are simple:

{Index - MIN(Index) / MAX(Index) - MIN(Index)} * 100

However, I'm having trouble implementing this as a measure that will calculate and change on the fly when the user updates the parameters. 

Here is the weighted linear combination as a measure: 

I tried using variables to yield the scores out of 100 for each row, but what I have currently yields NaN values.  

Any thoughts on how I might be able to get to the scores and keep them dynamic so they change when the parameters are adjusted?

Thanks

[Anonymous · Accepted solution]

Just following up on this with the solution. The issue is with the min/max calculations. MINX and MAXX are evaluated over every row in the data where the measure is calculated. They don't "anchor" to the max/min across all rows. Had to nest an all function: 

maxx(all(ta_vital),[TAVI_RAW]). Once you do that for min/max, the calculation works and the data is rescaled out of 100. 

[MisterFry]

Nothing else stands out as the possible cause of the issue. Your use case (in measures) is basically what what-if parameters are for.  

Probably not the issue... but one thing that jumps out at me is the use of SUMX and MAXX/MINX. Because you're just looking at a single value in the table SUM() and SUMX() should return the same thing, as would MINX/MAXX. 
What I would do at this stage, is test the return values of each of the VARs you've declared to narrow down the source of the issue. 

[MisterFry]

If you're creating this as a 'Column' in DAX, column evaluation is done once when the data is loaded. Changing a parameter will not cause the table to re-evaluate. 

Referring to parameter values in a column calculation will yield BLANK() values, which is why your calculations are returning NaNs. 

You may be able to implement this as a MEASURE instead, but it depends on what you plan to do with it. 

[Anonymous]

Yeah I did start with attemtping a calculated column for everything but like you said, columns added to the data will not update on the fly. I'm trying to do all of this using measures. I think part of the problem is telling PowerBI to duplicate the 0/100 scoring for each value on an axis. I can create some simple measures for Min/Max. The trouble starts when I try to perform the mathematics with those measures and the "i-th" value of the raw index.