Report metrics for missing data polishing

na_polish_metrics(before, after)

Arguments

before, after

Tsibbles before and after polishing.

Value

A tibble contains:

  • prop_na & nobs_na: The proportion of NAs in the sliced data (the difference between before and after).

  • prop_removed, nobs_removed, nrows_removed, & ncols_removed: The proportion of removed observations over the overall observations.

Details

The metric used for measuring the effect of polishing events is prop_na * prop_removed. We'd like to minimise the loss by minimising both prop_na and prop_removed over sequential polishing events.

See also

Other missing value polishing functions: na_polish_auto, na_polish_measures