USEPA
Assign different censored data handling methods to different characteristics
Is your feature request related to a problem? Please describe:
At Cybertown 2025, we received the following related feedback: "It would be helpful if we could treat censored data differently by Parameter. This is probably too complicated to include. We are just going to have to figure out how to change data on our end to do this as some parameters we would want to use the detection limit and others we would want to almost zero out the value."
Describe the solution you'd like:
It may be possible to add a characteristic name or comparable data ID function input to the simple censored data handling function. This would only apply the selected method to the selected characteristic(s). It could be run multiple times with different methods assigned for different characteristics.
Describe alternatives you've considered:
Needs more research on feasibly.
Reminders for TADA contributors addressing this issue:
New features and/or edits should include all the following work:
-
[ ] Create or edit the function/code.
-
[ ] Document all code using line/inline and/or multi-line/block comments to describe what is does.
-
[ ] Create or edit tests in tests/testthat folder to help prevent and/or troubleshoot potential future issues.
-
[ ] Create or edit the function documentation. Include working examples.
-
[ ] Update or add the new functionality to the appropriate vignette (or create new one).
-
[ ] If function/code edits made as part of this issue impact other functions in the package or functionality in the shiny app, ensure those are updated as well.
-
[ ] Run TADA_UpdateAllRefs(), TADA_UpdateExampleData(), styler::style_pkg(), devtools::document(), and devtools::check() and address any new notes or issues before creating a pull request.
-
[ ] Run more robust check for releases: devtools::check(manual = TRUE, remote = TRUE, incoming = TRUE)
I'd like to work on this. We can discuss priorities and when I should start on it.
Ideas on how this could work:
- User specified comparable Id(s) as a param in the function and then can run multiple times with different comparable ids and methods selected each time
- User supplies a ref df w/ each comparable Id and the corresponding OD and ND methods to apply, there would need to be a default method (do nothing?) for comparable Ids that were not included in the ref and maybe a print message to the user that not all comparable Ids were included so defaults were being applied to a subset.