R/make_concordance_df.R

In emilelatour/lamisc: Latour's Little Helpers

#' @title Make a table of concordance for multi-rater kappa
#'
#' @description
#' This function is a helper for working with the /code{concordance} function
#' from the \code{raters}. Rather than a table or the orginal columns of data,
#' the \code{concordance} function takes, as an input, a data frame that shows
#' for each thing being rated (patient, sample, etc.) the number of raters that
#' gave a particular rating. It creates a data frame of counts of raters by
#' ratee and rating.
#'
#' @param df A data frame or tibble.
#' @param id A column/ariable (any type) that uniquely identifies the unit being
#'   rated (i.e. patient id, ssn, name, etc.).
#' @param ... The columns of ratings by the raters, character or factor.
#' @param keep_id Logical; if /code{TRUE} then it will keep the /code{id} column
#'   else it will be dropped.
#'
#' @import dplyr
#' @import tidyr
#' @import rlang
#' @importFrom rlang .data
#'
#' @return A tibble
#' @export
#'
#' @examples
#' library(tibble)
#' library(raters)
#' diagnostic_df <-  data.frame(stringsAsFactors=FALSE,
#'          id = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L,
#'                 15L, 16L, 17L, 18L, 19L, 20L, 21L, 22L, 23L, 24L, 25L, 26L,
#'                 27L, 28L, 29L, 30L),
#' rater_1 = c("Neurosis", "Personality disorders", "Schizophrenia", "Other",
#'             "Personality disorders", "Depression", "Schizophrenia",
#'             "Depression", "Depression", "Other", "Depression", "Depression",
#'             "Personality disorders", "Depression", "Personality disorders",
#'             "Schizophrenia", "Depression", "Depression", "Personality disorders",
#'             "Depression", "Other", "Neurosis", "Personality disorders",
#'             "Depression", "Depression", "Personality disorders", "Depression",
#'             "Personality disorders", "Depression", "Other"),
#' rater_2 = c("Neurosis", "Personality disorders", "Schizophrenia", "Other",
#'             "Personality disorders", "Depression", "Schizophrenia",
#'             "Depression", "Depression", "Other", "Neurosis",
#'             "Personality disorders", "Personality disorders", "Neurosis",
#'             "Personality disorders", "Schizophrenia", "Depression", "Depression",
#'             "Personality disorders", "Schizophrenia", "Other", "Neurosis",
#'             "Personality disorders", "Depression", "Neurosis", "Personality disorders",
#'             "Depression", "Personality disorders", "Schizophrenia", "Other"),
#' rater_3 = c("Neurosis", "Personality disorders", "Schizophrenia", "Other",
#'             "Personality disorders", "Schizophrenia", "Schizophrenia",
#'             "Schizophrenia", "Neurosis", "Other", "Neurosis", "Neurosis",
#'             "Personality disorders", "Neurosis", "Neurosis", "Schizophrenia",
#'             "Depression", "Depression", "Neurosis", "Schizophrenia", "Other",
#'             "Neurosis", "Neurosis", "Neurosis", "Neurosis",
#'             "Personality disorders", "Depression", "Neurosis", "Schizophrenia", "Other"),
#' rater_4 = c("Neurosis", "Other", "Schizophrenia", "Other", "Neurosis",
#'             "Schizophrenia", "Schizophrenia", "Schizophrenia", "Neurosis",
#'             "Other", "Neurosis", "Neurosis", "Schizophrenia", "Neurosis",
#'             "Neurosis", "Schizophrenia", "Neurosis", "Depression",
#'             "Neurosis", "Other", "Other", "Neurosis", "Other", "Neurosis", "Neurosis",
#'             "Personality disorders", "Depression", "Neurosis",
#'             "Schizophrenia", "Other"),
#' rater_5 = c("Neurosis", "Other", "Personality disorders", "Other",
#'             "Neurosis", "Schizophrenia", "Other", "Schizophrenia",
#'             "Neurosis", "Other", "Neurosis", "Neurosis", "Schizophrenia",
#'             "Neurosis", "Neurosis", "Schizophrenia", "Other", "Depression",
#'             "Neurosis", "Other", "Other", "Neurosis", "Other", "Neurosis",
#'             "Neurosis", "Personality disorders", "Other", "Neurosis", "Schizophrenia",
#'             "Other"),
#' rater_6 = c("Neurosis", "Other", "Other", "Other", "Neurosis",
#'             "Schizophrenia", "Other", "Neurosis", "Neurosis", "Other",
#'             "Neurosis", "Neurosis", "Schizophrenia", "Neurosis", "Other",
#'             "Other", "Other", "Personality disorders", "Neurosis", "Other",
#'             "Other", "Personality disorders", "Other", "Neurosis", "Other",
#'             "Neurosis", "Other", "Neurosis", "Schizophrenia", "Other")
#' )
#'
#' diagnostic_2 <- make_concordance_df(df = diagnostic_df,
#'                                   id = id,
#'                                   rater_1:rater_6,
#'                                   keep_id = FALSE)
#' diagnostic_2
#' data("diagnostic")
#' raters::concordance(diagnostic, test = "Chisq")
#' raters::concordance(diagnostic_2, test = "Chisq")

make_concordance_df <- function(df, id, ..., keep_id = TRUE) {

  id <- rlang::enquo(id)
  vars <- rlang::enquos(...)

  res <- df %>%
    dplyr::select(!! id, !!! vars) %>%
    tidyr::gather(key = "key", value = "value", !!! vars) %>%
    dplyr::count(!! id, .data$value) %>%
    tidyr::spread(., key = .data$value, value = .data$n) %>%
    dplyr::mutate_all(.tbl = .,
                      .funs = funs(tidyr::replace_na(., 0)))

  if (keep_id) {
    res
  } else {
    res %>%
      dplyr::select(-!! id)
  }

}