How to detect outliers

[#1]

Hi all,

I would like to make a script to detect outliers based on my data population.

Based on the image below, I tried to calculate mean of all pockets from 0 to 6 and applied method ((Mean +- (3*sigma)) to find outliers. By direct observation, Pocket 2 has a lot of outliers. However, in case the value of variation data points on Pocket2 is within ((Mean+-(3*sigma)), my method cannot detect them as outliers. 

If I use method ((Mean +-(2*sigma)) to detect outliers, it is too sensitive for other cases.

Does anyone have any advice for me

Sincerely,

[#2]

These might give you some ideasExploring Outliers in Your Data and Outliers Episode 1: The elusive outlier described, visually identified, and judged