mists-polish: Missing data polishing for tsibble
In earowang/mists: Missingness of Temporal Data

Description Usage Arguments Details See Also Examples

If the proportion of overall missings is less than the cutoff,

na_polish_measures() removes columns or observations.
na_polish_key() polishes data by rows or observations, removing the whole rows of key series.
na_polish_index() polishes data by rows or observations, removing either the starting or the ending NA blocks (if any) within each key series.
na_polish_index2() polishes data by rows or observations, removing the ending NA blocks (if any) within each key series. It is a shortcut of na_polish_index(na_fun = na_ends_with).

na_polish_measures(data, cutoff)

na_polish_key(data, cutoff)

na_polish_index(data, cutoff, na_fun = na_starts_with)

na_polish_index2(data, cutoff)

`data`	A tsibble.
`cutoff`	A numeric between 0 and 1. Rows/cols will be kept, if the proportion of overall missings is less than the cutoff.
`na_fun`	Either `na_starts_with` or `na_ends_with`.

The proportion of overall missings is defined as the number of NA divided by the number of measurements (i.e. excluding key and index).

Other missing value polishing functions: na_polish_auto, na_polish_metrics

wdi_ts <- tsibble::as_tsibble(wdi, key = country_code, index = year)
(wdi_cols <- na_polish_measures(wdi_ts, cutoff = .7))
# columns removed
setdiff(names(wdi_ts), names(wdi_cols))
na_polish_key(wdi_ts, cutoff = .7)
na_polish_index(wdi_ts, cutoff = .7)
na_polish_index2(wdi_ts, cutoff = .7)