Title: Visualize Contingency Tables Using Kite-Square Plots
Version: 0.0.2
Description: Create a kite-square plot for contingency tables using 'ggplot2', to display their relevant quantities in a single figure (marginal, conditional, expected, observed, chi-squared). The plot resembles a flying kite inside a square if the variables are independent, and deviates from this the more dependence exists.
License: LGPL (≥ 3)
Encoding: UTF-8
URL: https://github.com/HUGLeipzig/kitesquare
BugReports: https://github.com/HUGLeipzig/kitesquare/issues
RoxygenNote: 7.3.2
VignetteBuilder: quarto
Imports: tidyr, dplyr, ggplot2, ggh4x, scales, rlang
Suggests: knitr, quarto
Config/Needs/website: quarto
NeedsCompilation: no
Packaged: 2025-03-19 12:42:45 UTC; Wiedenhoef
Author: John Wiedenhöft ORCID iD [aut, cre, cph]
Maintainer: John Wiedenhöft <john.wiedenhoeft@medizin.uni-leipzig.de>
Repository: CRAN
Date/Publication: 2025-03-20 13:50:09 UTC

Create a kite-square plot

Description

Given a data frame or tibble, creates a kite-square plot to visualize the contingency table of two discrete variables.

Usage

kitesquare(
  df,
  x,
  y,
  obs,
  normalize = FALSE,
  full_range = FALSE,
  center_x = TRUE,
  center_y = TRUE,
  center = NULL,
  fill_x = FALSE,
  fill_y = FALSE,
  fill = NULL,
  kite = TRUE,
  spars = TRUE,
  square = TRUE,
  chi2 = TRUE,
  bars_x = TRUE,
  bars_y = TRUE,
  bars = NULL,
  intersect_x = TRUE,
  intersect_y = TRUE,
  intersect = NULL,
  color_x = "#e31a1c",
  color_y = "#1f78b4",
  kite_color = "black",
  square_color = "black",
  spars_color = "black",
  chi2_color = "#bebebe",
  border_color = "black",
  alpha_fill = 0.3,
  alpha_chi2 = 0.3,
  alpha = NULL,
  pointsize = 3,
  linewidth = 1,
  whisker_length = 0.05,
  extend_whiskers = FALSE,
  dodge_x = 2,
  ...
)

Arguments

df

A tibble or data frame of observations.

x, y

Name of the variable in df for x (columns) and y (rows), as string or expression.

obs

Name of observation counts in df. If a combination of y and x appears multiple times in df, obs are added together. If not provided, a value of 1 will be assumed for each line in df.

normalize

Should values te normalized to probabilities and expressed in percent?

full_range

If normalize is TRUE, should all axes limits be from 0 to 1?

center_x, center_y, center

Should a binary x or y variable be centered (axis reversed) so that the spars meet? center overrides both.

fill_x, fill_y, fill

Should the space between x or y bars and the axis be filled? fill overrides both.

kite, square, spars, chi2

Should the kite, square, spars and \chi^2 patches be drawn?

bars_x, bars_y, bars

Should the bars for the x and y variables be drawn? bars overrides both.

intersect_x, intersect_y, intersect

Should the intersect positions for x and y variables with their axes be drawn? intersect overrides both.

color_x, color_y

Colors for x and y.

kite_color, square_color, spars_color, chi2_color

Color of the kite, square, spars and \chi^2 patches.

border_color

Color for the border around each cell.

alpha_fill, alpha_chi2, alpha

Transparency for fill and \chi^2 patches. alpha overrides both.

pointsize

The point size for intersects and spars.

linewidth

The line width for bars and spars.

whisker_length

The length of bar whiskers.

extend_whiskers

Should the bar whiskers be extended to wrap around the \chi^2 patches?

dodge_x

The number of levels the x axis labels should dodge.

...

Further arguments passed to ggplot2::facet_grid().

Value

A ggplot object, with an extra $table key. The latter contains the tibble of coordinates created internally for plotting.

Examples

df <- dplyr::tibble(
  X=c('A', 'A', 'B', 'B', 'B'),
  Y=c('U', 'V', 'U', 'V', 'V'),
  count=c(30,15,30,70,65))
kitesquare(df, X, Y, count)
kitesquare(df, X, Y, count, normalize=TRUE, center_x=FALSE)