Outlier detection globally

[MMocz]

I'd like to detect outliers using the interquartile method, but I've only found this post:
Calculate Quartiles and Interquartile Range to Detect Outliers - Tips & Tricks - Seeq Community Website

When I create the quartiles I need to set bounding condition, but I'd like to use the global time-range to get the quartiles. (not periodically)

Does anyone have information on how to calculate quartiles, interquartiles globally (full-analysed-time-range) in seeq?

 

[Chris Harp]

Hi MMocz,

Within Seeq Workbench, you cannot use the current display range or the investigation window to define a capsule.  There are other options such as creating a sliding window using the now() function to define a condition or create a large capsule that will take into account past and future data.

Example: 

condition(7d, capsule(now()-7d, now()))

Example 2:

condition(4y, 
capsule('2022-05-29T00:00-05:00', '2026-05-29T00:00-05:00'))
// Optional to stop capsule at now
//.intersect(past())

When creating signals from these conditions, the data will be uncertain and will not cached.  Large analyses may result in long load times since the signals must be recalculated each time data is refreshed.  

Thanks,

Chris