Measure to group and sum - perform calculation on subgroups - sum subgroups

[JS_pbi]

I am trying to create a dynamic power curve regression but I need to do the following to generate the c and b of y=c*x^b which require the following steps:

So, can I do the following based on the  explanatory data below?

1. Group by X and sum Y
2. Extract the natural log from these aggregated subgroups of Y using LN function
3. Sum the results of step 2

Expected Result = -41.93

There are other columns/dimensions which I will filter on and have the measure result dynamically change with the formula.

I have spent many hours researching this and all I have come to is the following formula which only generates the natural log of the sum of all values in the field - not the subsets.   

    Measure = CALCULATE(LN(SUM('Power Curve Test Data'[y])),'Power Curve Test Data'[x])

X              Y

0	0.16
1	0.1
2	0.07
3	0.06
4	0.058
5	0.052
6	0.042
7	0.041
8	0.038
9	0.032
10	0.029
11	0.027
12	0.022
13	0.019
14	0.012
0	0.116517
1	0.099084
2	0.027853
3	0.057962
4	0.025471
5	0.00566
6	0.017254
7	0.03229
8	0.0083
9	0.029146
10	0.00775
11	0.005638
12	0.019069
13	0.000277
14	0.005248

[v-lili6-msft · Accepted solution]

Hi @ JS_pbi You can use SUMMARIZE function AND ADDCOLUMNS function Try this formula please

Measure = SUMX(ADDCOLUMNS(SUMMARIZE('Power Curve Test Data','Power Curve Test Data'[X],"A",SUM('Power Curve Test Data'[Y])), "ln", ln([A])), [ln])

Result:

Best Regards,

Lin

[Anonymous]

can you explain this formula, I understood untill SUM part but after that where do "ln" and in([A]) came from 

This would be nice if you explain whole formula to make it easy to understand

Thanks

[JS_pbi]

Thank You!