Make crosstabs and add margin totals (or means); a tidyverse-friendly version of base::table()

margintable.R

#' Crosstabs with margins
#'
#' @param data Data frame or tibble
#' @param row Column name (string or index) for rows
#' @param col Column name (string or index) for columns
#' @param margin_func Margin function (e.g., sum, mean, median)
#' @param margin_label Label for margin column and row 
#' @param col_prefix Prefix for columns (defaults to col name), set to FALSE to omit; If return == "kable", this is used for the grouping header
#' @param return Return a tibble or formatted kableExtra table
#' @param ... Arguments to pass to kableExtra::kable()
#'
#' @return tibble or kableExtra table
#' @export
#'
#' @examples
#' margintable(mtcars, "cyl", "vs")
#' margintable(mtcars, 2, 8)
#' margintable(mtcars, "vs", "cyl", col_prefix = FALSE)
#' margintable(mtcars, "vs", "cyl", mean, "Mean")
#' margintable(mtcars, "vs", "cyl", col_prefix = "Number of cylinders", return = "kable")
#' margintable(mtcars, "vs", "cyl", col_prefix = FALSE, return = "kable", format = "latex")
margintable <- function(data, row, col, 
                        margin_func = sum, 
                        margin_label = "Total",
                        col_prefix = col,
                        return = c("tibble", "kable"),
                        ...) {
  # convert numeric specification to names
  if (is.numeric(row)) row <- names(data)[[row]]
  if (is.numeric(col)) col <- names(data)[[col]]
  
  mt <- data %>%
    dplyr::count(.data[[row]], .data[[col]]) %>%
    tidyr::pivot_wider(names_from = dplyr::all_of(row),
                       values_from = n,
                       values_fill = 0) %>%
    dplyr::rowwise(dplyr::all_of(col)) %>%
    dplyr::mutate(!!margin_label := 
                  margin_func(dplyr::c_across())) %>%
    tidyr::pivot_longer(cols = -dplyr::all_of(col),
                        names_to = row,
                        values_to = "n") %>%
    tidyr::pivot_wider(names_from = dplyr::all_of(col),
                       values_from = n) %>%
    dplyr::rowwise(dplyr::all_of(row)) %>%
    dplyr::mutate(!!margin_label := 
                  margin_func(dplyr::c_across()))
  
  if (match.arg(return) == "tibble") {
    # optionally add column prefix
    if (!isFALSE(col_prefix)) {
      mt <- mt %>%
        dplyr::rename_with(.fn = ~paste0(col_prefix, "_", .x),
                           .cols = -c(1, ncol(.)))
    }
    return(mt)
  } else {
    headers <- setNames(object = c(1, ncol(mt)-2, 1),
                        nm = c(" ", col_prefix, " "))
    kt <- kableExtra::kable(mt, ...)
    if (!isFALSE(col_prefix)) {
      kt <- kt %>% kableExtra::add_header_above(headers)
    }
    return(kt)
  }
}