R/na_count.R

In misc: Miscellaneous Functions for Data and Geospatial Work

#' Count NA frequency in data
#'
#' @description
#' `na_count()` is a way to display the count and frequency of NA in data. It
#' can be slow over large datasets.
#'
#' @param data a data frame
#' @param sort If `TRUE`, sort rows by descending `na_percent`. If `FALSE`, keep
#'   column order from the input.
#'
#' @return a long-format tibble
#'
#' @importFrom dplyr summarise_all n full_join arrange desc
#' @importFrom rlang .data
#' @importFrom purrr map_df
#' @importFrom tidyr pivot_longer everything
#'
#'
#' @section Acknowledgment:
#' I learned this way of exploring data though the excellent webinar taught by
#' Emily Robinson.
#'
#' @export
#'
#' @examples
#' \donttest{
#' na_data <- data.frame(c1 = c(1, NA), c2 = c(NA, NA))
#' na_data %>% na_count()
#' }
na_count <- function(data, sort = TRUE) {
  na_count_data <-
    data %>%
    purrr::map_df(~ sum(is.na(.))) %>%
    tidyr::pivot_longer(
      cols = tidyr::everything(),
      names_to = "variables",
      values_to = "na_count"
    )
  na_percent_data <-
    data %>%
    dplyr::summarise_all(~ round((sum(is.na(.)) / dplyr::n()) * 100, digits = 3)) %>%
    tidyr::pivot_longer(
      cols = tidyr::everything(),
      names_to = "variables",
      values_to = "na_percent"
    )

  full_check <- dplyr::full_join(x = na_count_data, y = na_percent_data, by = "variables")

  return({
    if (sort) {
      full_check %>%
        dplyr::arrange(dplyr::desc(.data$na_percent))
    } else {
      full_check
    }
  })
}