marimekko: marimekko and Mosaic Plots for 'ggplot2'

Description

Create marimekko (mosaic) plots as a 'ggplot2' layer. Column widths encode marginal proportions of one categorical variable and segment heights encode conditional proportions of a second categorical variable.

Author(s)

Maintainer: Dawid Kaledkowski dawid.kaledkowski@gmail.com

Retrieve computed tile positions from a marimekko layer

Description

Retrieve computed tile positions from a marimekko layer

Usage

StatMarimekkoTiles

Format

An object of class StatMarimekkoTiles (inherits from Stat, ggproto, gg) of length 2.

Usage with custom geoms

Use StatMarimekkoTiles as the stat argument in ggplot2::layer() to pair the tile data with any geom. The only requirement is that geom_marimekko() must appear before the custom layer so that tile positions are computed first.

See Also

geom_marimekko(), geom_marimekko_text(), geom_marimekko_label(), fortify_marimekko()

Examples

library(ggplot2)

titanic <- as.data.frame(Titanic)

# Bubble overlay — point size encodes tile count
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, weight = Freq),
    formula = ~ Class | Survived, alpha = 0.4
  ) +
  layer(
    stat = StatMarimekkoTiles,
    geom = GeomPoint,
    mapping = aes(size = after_stat(weight)),
    data = titanic,
    position = "identity",
    show.legend = FALSE,
    inherit.aes = FALSE,
    params = list(colour = "white", alpha = 0.7)
  ) +
  scale_size_area(max_size = 12)

# Residual markers — colour and size show deviation from independence
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, weight = Freq),
    formula = ~ Class | Survived
  ) +
  layer(
    stat = StatMarimekkoTiles,
    geom = GeomPoint,
    mapping = aes(
      size = after_stat(abs(.residuals)),
      colour = after_stat(ifelse(.residuals > 0, "over", "under"))
    ),
    data = titanic,
    position = "identity",
    show.legend = TRUE,
    inherit.aes = FALSE,
    params = list(alpha = 0.8)
  ) +
  scale_colour_manual(
    values = c(over = "tomato", under = "steelblue"),
    name = "Deviation"
  ) +
  scale_size_continuous(range = c(1, 8), name = "|Residual|")

Compute marimekko tile rectangles as a data frame

Description

Compute marimekko tile rectangles as a data frame

Usage

fortify_marimekko(
  data,
  formula,
  weight = NULL,
  gap = 0.01,
  gap_x = NULL,
  gap_y = NULL,
  standardize = FALSE
)

Arguments

Value

A data frame with columns for each formula variable, plus fill, colour, xmin, xmax, ymin, ymax, x, y, weight, .proportion, .marginal, and .residuals.

Examples

titanic <- as.data.frame(Titanic)
fortify_marimekko(titanic, formula = ~ Class | Survived, weight = Freq)

# 3-variable formula
fortify_marimekko(titanic, formula = ~ Class | Survived | Sex, weight = Freq)

Generalized mosaic plot with formula-based variable nesting

Description

Generalized mosaic plot with formula-based variable nesting

Usage

geom_marimekko(
  mapping = NULL,
  data = NULL,
  formula = NULL,
  gap = 0.01,
  gap_x = NULL,
  gap_y = NULL,
  colour = NULL,
  alpha = 0.9,
  show_percentages = FALSE,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE,
  ...
)

Arguments

Value

A list of ggplot2 layers (geom + axis scales).

How the formula works

The formula uses two operators to encode the full partitioning hierarchy in a single expression:

| (pipe)

Separates nesting levels. Each | switches the splitting direction, alternating horizontal, vertical, horizontal, vertical, and so on. The first variable (or group) listed is the outermost split — it partitions the entire plot area. Each subsequent level partitions the tiles created by the previous level.

+ (plus)

Groups variables at the same nesting level. All variables joined by + share the same splitting direction and are applied sequentially within that level. The first + variable partitions the current tiles, then the second + variable further subdivides those tiles, still in the same direction.

Reading order — outermost to innermost

The formula is read left to right, from the coarsest (outermost) partition to the finest (innermost):

~ a | b

First split the plot horizontally by a (columns whose widths reflect marginal proportions of a). Then, within each column, split vertically by b (rows whose heights reflect conditional proportions of b given a). This is the classic two-variable marimekko / mosaic plot.

~ a | b | c

Horizontal by a, then vertical by b, then horizontal again by c. Three levels of nesting with alternating directions (h \to v \to h).

~ a + b | c

Horizontal by a, then horizontal again by b (same direction because + groups them), then vertical by c. This is the double decker pattern — all horizontal splits first, with a single vertical split at the end.

~ a | b + c

Horizontal by a, then vertical by b, then vertical again by c. Two vertical variables nested within each column.

Computed variables

The stat computes the following variables that can be accessed with ggplot2::after_stat():

.proportion

Conditional proportion of the tile within its immediate parent. For a formula ~ a | b, this is the proportion of b within each level of a, i.e. P(b \mid a). Values sum to 1 within each parent tile. Useful for mapping to alpha to fade tiles by their local share: aes(alpha = after_stat(.proportion)).

.marginal

Joint (marginal) proportion of the tile relative to the whole dataset, i.e. n_\text{cell} / N. Values sum to 1 across all tiles. Used internally for x-axis percentage labels when show_percentages = TRUE, and can be mapped to aesthetics to emphasise cells by overall frequency.

.residuals

Pearson residual measuring departure from statistical independence between the horizontal and vertical variable groups. Computed as (O - E) / \sqrt{E}, where O is the observed cell count and E is the count expected under independence. Positive values indicate the cell is more frequent than expected; negative values indicate less frequent. When only one direction (all horizontal or all vertical) is present, .residuals is set to 0. Map to alpha or fill to highlight deviations: aes(alpha = after_stat(abs(.residuals))).

Examples

library(ggplot2)

titanic <- as.data.frame(Titanic)

# 2-variable mosaic
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, weight = Freq),
    formula = ~ Class | Survived
  )

# 3-variable mosaic (h -> v -> h)
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, weight = Freq),
    formula = ~ Class | Survived | Sex
  )

# Multi-variable fill with interaction()
ggplot(titanic) +
  geom_marimekko(
    aes(fill = interaction(Sex, Survived), weight = Freq),
    formula = ~ Class | Sex + Survived
  )

# Fade tiles by conditional proportion
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, alpha = after_stat(.proportion), weight = Freq),
    formula = ~ Class | Survived
  ) +
  guides(alpha = "none")

# Highlight cells that deviate from independence
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, alpha = after_stat(abs(.residuals)), weight = Freq),
    formula = ~ Class | Survived
  ) +
  guides(alpha = "none")

Add labels with background to a marimekko plot

Description

Add labels with background to a marimekko plot

Usage

geom_marimekko_label(
  mapping = NULL,
  data = NULL,
  position = "identity",
  ...,
  size = 3.5,
  colour = "black",
  fill = alpha("white", 0.7),
  label.padding = unit(0.15, "lines"),
  na.rm = FALSE,
  show.legend = FALSE,
  inherit.aes = FALSE
)

Arguments

Value

A ggplot2 layer.

Examples

library(ggplot2)

titanic <- as.data.frame(Titanic)
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, weight = Freq),
    formula = ~ Class | Survived
  ) +
  geom_marimekko_label(aes(label = after_stat(weight)))

Add text labels to a marimekko plot

Description

Add text labels to a marimekko plot

Usage

geom_marimekko_text(
  mapping = NULL,
  data = NULL,
  position = "identity",
  ...,
  size = 3.5,
  colour = "white",
  na.rm = FALSE,
  show.legend = FALSE,
  inherit.aes = FALSE
)

Arguments

Value

A ggplot2 layer.

Examples

library(ggplot2)

titanic <- as.data.frame(Titanic)
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, weight = Freq),
    formula = ~ Class | Survived
  ) +
  geom_marimekko_text(aes(label = after_stat(weight)))

Unikko-inspired colour palette

Description

A character vector of 8 bold colours inspired by Marimekko's iconic Unikko poppy pattern. Vibrant, high-contrast tones suited for categorical data visualisation.

Usage

marimekko_pal

Format

An object of class character of length 8.

Minimal theme for marimekko plots

Description

Removes x-axis gridlines and adjusts spacing for mosaic plots. Also applies the marimekko_pal fill scale.

Usage

theme_marimekko(base_size = 12, ...)

Arguments

Value

A ggplot2 theme.

Examples

library(ggplot2)

titanic <- as.data.frame(Titanic)
ggplot(titanic) +
  geom_marimekko(
    aes(fill = Survived, weight = Freq),
    formula = ~ Class | Survived
  ) +
  theme_marimekko()