Thank you.
I got my typos, and fixed it.
530 -531 = -1 (instead of 1)
It will mean that nothing is potential outlier in below Q1 range now.
As I stated above, I have intention to truncate extreme values, outliers for later analysis.
From real dataset, there are 24 potential outliers (nearly 8.4% of 286 observed days so far), which listed below. Only three of those 24 outliers occured after the first of Mar. In other words, 87.5% (21/24) of those outliers occured within January to February.
Full dataset will be used only for descriptive plots, whilst truncated one will be used for analysis.
Extreme values or outliers dramatically affect on results, and might mislead readers far away from the truth.
I think, you have to leave some datapoints in doing this analysis like start of Merit system and the day when number of Merit source increased. These datapoint are affected and are not same as their previous day.