Bar plot for a set of multiple-choice items — stem

Plots a set of "select all that apply" items, showing the share of respondents that picked each option. The items are reshaped to long format and their (possibly weighted) frequencies are rescaled by the number of items so that the bars express percentages of respondents rather than of responses.

Usage

stem_multiselect(
  data,
  items,
  group = NULL,
  weight = NULL,
  palette = "div1",
  direction = 1,
  labels = TRUE,
  label_accuracy = 1,
  label_suffix = "",
  label_hide = 0,
  infreq_order = TRUE,
  background = TRUE,
  background_fill = "grey85",
  background_alpha = 1
)

Arguments

data: Data frame holding the item (and group) variables.
items: Items to plot. Supports tidyselect helpers (e.g. dplyr::starts_with()).
group: Optional grouping variable. When supplied, one stacked horizontal bar is drawn per group category (with the item mapped to fill) and proportions are computed within each group.
weight: Optional survey weights.
palette: Name of a Stem palette. See stem_palette().
direction: Palette direction. Use -1 to reverse the colours.
labels: If TRUE, prints a percentage label on each bar.
label_accuracy: Rounding accuracy of labels. 1 gives whole numbers, 0.1 one decimal place.
label_suffix: Suffix appended to labels. Defaults to "".
label_hide: Proportions below this threshold are left unlabelled. Defaults to 0.05 when a group is supplied (to keep small stacked segments unlabelled) and 0 otherwise.
infreq_order: If TRUE (default), options are ordered by frequency.
background: If TRUE (default, ungrouped plots only), draws a full-width background bar behind each option.
background_fill: Fill colour of the background bars.
background_alpha: Opacity of the background bars.

Value

A ggplot2 object.

Examples

if (FALSE) { # \dontrun{
stem_multiselect(trust, items = dplyr::starts_with("biggest_concern"))
stem_multiselect(trust, items = dplyr::starts_with("biggest_concern"), weight = W)
} # }