'Morning!
Stdevp as reference line was used to simulate the behavior
The goal is the following:
To have a bar chart with the distinct count of patient IDs, by month (x axis), by hospital (bar colors). The bar chart should display a reference line with the stdevp of distinct count by month, ignoring the subdimension.
This is a sample of the dataset:
And this is the indended behavior:
The problem has been that I can’t figure out a way to calculate the value ignoring the ‘Hospital’ dimension.
With the formula
[Reference Line Value]
stdevpOver(distinct_count({Patient ID}), [])
I get the following behavior. In the third table the values change, because now the sample size is bigger (since as a table calculation it will consider all ‘sub-values’ as well).
And that is also reflected when trying to plot the values. By creating reference lines with max, min and average we can see there are multiple values, which doesn’t happen with the first one.
I have already tried a mix of combinations of LAC-As and -Ws, and corresponding dimensions, but still can’t figure out a way to make it work.