I have a scatterplot where the x-axis is a paid dollar amount and the y-axis is a count and each dot represents a different person. Both fields are calculated measures. I have used the scatterplot visual to create lines on the chart for the mean, 68th percentile (1SD), and 95th percentile (2SD) for each axis. I have multiple filters at the top such as date range, location, etc. The mean and standard deviations would change based on the filter selections. I want to color code the points on the scatterplot to be a different color if both axis' are >= 2SD (how we are defining an outlier). I've seen outlier detection methods for PowerBI but they seem to be based on more static thesholds, while mine would not be.
I was able to create another measure field for the SD's (since the visual automatically calculated it) and I thought I needed a calculated field so I could just compare that in a formula. But I want to compare each dot/person's count to the total standard deviation of the count (and also each person's paid dollar amount to the standard deviation of that), which I'm unable to compute. Referencing the table below for example, I'd want to compare the count field for each person to the total of the std_of_count field and that would be an outlier, but instead I'm only able to figure out how to do it for each row. I believe this would be a start to figuring it out so I can label this as an outlier (1) or not (0), then I can use the data colors/markers field within the visual to color code this as a rule.
| PERSON | COUNT (y-axis) | STD_OF_COUNT |
| Person 1 | 5 | 5 |
| Person 2 | 3 | 3 |
| ... | ... | ... |
| Person 50 | 1 | 1 |
| TOTAL | 200 | 1 |
