Can someone help me to conclude on how this 28.3% is calculated?
I can think of the following options:
a. 28.3% more users on the last 1 hour (based on Step)
b. 28.3% more users based on the same date 7 days ago (based on query option relative time)
c. 28.3% more users based on the last data point (we have one datapoint every 10 minutes)
“c” does not makes a lot of sense, as my app does not behave as such.
I had a quick look at the code, and I think the relevant code bit is right here:
The way the calculation works is as follows:
The field gets reduced through the use of a special field reducer in the transformations package, called “diffperc”
The way that calculation works is that it reuses the “diff” calculation, and then divides by the first non-null field value.
The way “diff” works is that it subtracts the first non-null from the last non-null value. (See reference here)
So over a 7-day period as in your screenshot, I think what’s going on here is that the first value is subtracted by the last value, then divided by the “starting value” (first non-null field value) and you get the result like that. Example, let’s say your data is:
null, null, 1, 2, 3, 4, 5, null, 5, 4, null
Diff would be 4 - 1, and diffperc would be (4-1) / 1 = 300%
