I am looking to hide and exclude the top and bottom 0.5% of data per identification number. For example, the 99.5% and 0.5% quantiles would be hidden and excluded for ID #123, then for ID #456, #789. Each ID may have a unique range of values, so I do not want to group all the ID's together and remove what could be valuable data. There are also a reasonable number of ID's, so I am hoping for a somewhat efficient way to complete this task--either scripting or interactive explanation is fine. Please send help! Thanks!
You could set up three columns. The first column would compute the low quantile. The second column would compute the high quantile. The third column would indicate if the observation is beyond either quantile.