evidently icon indicating copy to clipboard operation
evidently copied to clipboard

Add relative percent change to TestColumnShareOfMissingValues

Open elutins opened this issue 1 year ago • 2 comments

Add relative percent change to TestColumnShareOfMissingValues

I want to be able to make a test pass/fail if the relative percent change of null values exceeds a certain threshold. I understand that the default behavior for the test is to fail if the share of missing values in a column is over 10% higher than in reference. as stated in this section of the doc. I would ideally like to change this 10% value to something else. I understand the lt parameter can be set, however that set the absolute threshold with no relation back to the reference dataset. Is there a way to do this?

Here is how I am currently using/creating the test:

tests = []
tests.append(
   generate_column_tests(
     TestColumnShareOfMissingValues,
     columns="all",
     parameters={"lt": missing_values_threshold},
  )
)
test_suite = TestSuite(tests=suite_tests)

Instead of using the lt parameter in the above code snippet, Id ideally like to pass a different parameter that sets the relative percent change threshold from the reference to the current dataset.

elutins avatar Feb 16 '24 21:02 elutins

Hi, @elutins - it is a great feature request, and we'd like to implement similar functionality eventually. However, right now, it is not possible to set a custom condition relative to the reference.

You can either:

  • use the default auto-generated relative test conditions
  • set custom absolute conditions
  • set custom conditions relative to an absolute value using approx https://docs.evidentlyai.com/user-guide/tests-and-reports/custom-test-suite#custom-conditions-with-approx (you still need to set the value you are referring to).

elenasamuylova avatar Feb 16 '24 22:02 elenasamuylova

sounds good - thanks for the update. would be a great feature

elutins avatar Feb 16 '24 22:02 elutins