Comparison with patchwork and cowplot

Introduction

R offers several packages for arranging multiple plots in a single figure. This vignette compares three popular approaches:

customLayout - matrix-based layout specification with aspect ratio preservation
patchwork - operator-based plot composition for ggplot2
cowplot - publication-ready figure composition for ggplot2

Each package has its strengths and is suited for different use cases.

Feature Comparison Table

Feature	customLayout	patchwork	cowplot
Base graphics support	Yes	No	No
ggplot2 support	Yes	Yes	Yes
PowerPoint export	Yes	No	No
Matrix-based specification	Yes	No	No
Operator syntax (+, /, \|)	No	Yes	No
Auto panel labels	No	Yes	Yes
Legend collection	Manual	Yes	Manual
Inset plots	No	Yes	Yes
Aspect ratio preservation	Yes	Yes	Yes
Axis alignment	N/A	Yes	Yes

Package Overview

customLayout

customLayout extends R’s base layout() function to work with base graphics, grid graphics (ggplot2), and PowerPoint slides. It uses matrix-based layout specification and preserves aspect ratios when combining layouts.

patchwork

patchwork provides an intuitive operator-based syntax (+, /, |) for combining ggplot2 plots. It’s designed for simplicity and ease of use with ggplot2.

cowplot

cowplot offers publication-ready themes and tools for combining ggplot2 plots. It provides fine-grained control over plot alignment and annotation.

Basic Comparison

Let’s create the same layout using all three packages. We’ll arrange four ggplot2 plots in a 2x2 grid.

library(ggplot2)

# Create sample plots
set.seed(123)
df <- data.frame(
  x = rnorm(100),
  y = rnorm(100),
  group = sample(c("A", "B"), 100, replace = TRUE)
)

p1 <- ggplot(df, aes(x, y)) +
  geom_point() +
  ggtitle("Scatter Plot")

p2 <- ggplot(df, aes(x)) +
  geom_histogram(bins = 20) +
  ggtitle("Histogram of X")

p3 <- ggplot(df, aes(y)) +
  geom_histogram(bins = 20) +
  ggtitle("Histogram of Y")

p4 <- ggplot(df, aes(group, y)) +
  geom_boxplot() +
  ggtitle("Boxplot by Group")

Using customLayout

library(customLayout)

# Create a 2x2 layout matrix
lay <- lay_new(matrix(1:4, nrow = 2, byrow = TRUE))

# Display the layout structure
lay_show(lay)

# Arrange plots using the layout
lay_grid(list(p1, p2, p3, p4), lay)

Using patchwork

library(patchwork)

# Using operators: | for side-by-side, / for stacked
(p1 | p2) / (p3 | p4)

Using cowplot

library(cowplot)

# Using plot_grid
plot_grid(p1, p2, p3, p4, nrow = 2)

Complex Layouts

The packages differ more significantly when creating complex, non-uniform layouts.

Different Width Columns

Let’s create a layout where the left column is twice as wide as the right column.

customLayout

# Create two separate layouts and combine with width ratio
left <- lay_new(matrix(1:2, nrow = 2))
right <- lay_new(matrix(1:2, nrow = 2))
combined <- lay_bind_col(left, right, widths = c(2, 1))

lay_show(combined)

lay_grid(list(p1, p2, p3, p4), combined)

patchwork

# Use layout specification with widths
(p1 / p2) | (p3 / p4) + plot_layout(widths = c(2, 1))

cowplot

# Create separate columns and combine
left_col <- plot_grid(p1, p2, ncol = 1)
right_col <- plot_grid(p3, p4, ncol = 1)
plot_grid(left_col, right_col, rel_widths = c(2, 1))

Nested Layouts

Create a layout with one large plot on the left and three smaller plots stacked on the right.

customLayout

# Create nested layout
left <- lay_new(1)  # Single plot
right <- lay_new(matrix(1:3, nrow = 3))  # Three stacked plots
nested <- lay_bind_col(left, right, widths = c(2, 1))

lay_show(nested)

lay_grid(list(p1, p2, p3, p4), nested)

patchwork

p1 | (p2 / p3 / p4) + plot_layout(widths = c(2, 1))

cowplot

right_side <- plot_grid(p2, p3, p4, ncol = 1)
plot_grid(p1, right_side, rel_widths = c(2, 1))

Unique Features

customLayout: Base Graphics Support

Unlike patchwork and cowplot, customLayout works with base R graphics. Here’s a complex layout with 8 different plot types arranged in a sophisticated structure.

# Create a complex layout:
# - Top row: one wide plot spanning 2 columns, one narrow plot
# - Middle row: 3 equal plots
# - Bottom row: 2 plots on left, one tall plot on right spanning middle+bottom

# Build the layout by combining simpler layouts
top_row <- lay_new(matrix(c(1, 1, 2), nrow = 1))
middle_row <- lay_new(matrix(1:3, nrow = 1))
bottom_left <- lay_new(matrix(1:2, nrow = 1))

# Combine middle and bottom rows, then split field for the tall right plot
main_area <- lay_bind_row(middle_row, bottom_left, heights = c(1, 1))

# Add a sidebar on the right for plots 6-8 stacked vertically
sidebar <- lay_new(matrix(1:3, nrow = 3))
combined <- lay_bind_col(main_area, sidebar, widths = c(3, 1))

# Add top row
final_layout <- lay_bind_row(top_row, combined, heights = c(1, 2))

lay_show(final_layout)

par(mar = c(3, 3, 2, 1))
lay_set(final_layout)

set.seed(123)
x <- rnorm(100)
y <- x + rnorm(100, sd = 0.5)

# Plot 1: Time series (wide)
plot(cumsum(rnorm(200)), type = "l", col = "steelblue", lwd = 2,
     main = "Random Walk", xlab = "", ylab = "Value")
grid()

# Plot 2: Density plot (narrow)
plot(density(rnorm(500)), main = "Density", col = "darkgreen", lwd = 2)
polygon(density(rnorm(500)), col = rgb(0, 0.5, 0, 0.3))

# Plot 3: Scatter with regression
plot(x, y, pch = 19, col = rgb(0, 0, 1, 0.5), main = "Scatter + Regression")
abline(lm(y ~ x), col = "red", lwd = 2)

# Plot 4: Histogram
hist(rnorm(300), breaks = 25, col = "coral", border = "white", main = "Histogram")

# Plot 5: Boxplot comparison
boxplot(rnorm(50), rnorm(50, mean = 1), rnorm(50, mean = -0.5),
        col = c("lightblue", "lightgreen", "lightyellow"),
        names = c("A", "B", "C"), main = "Comparison")

# Plot 6: Bar plot
barplot(sort(table(sample(LETTERS[1:6], 100, replace = TRUE)), decreasing = TRUE),
        col = rainbow(6), main = "Frequencies")

# Plot 7: Pie chart
pie(c(35, 25, 20, 15, 5), labels = c("A", "B", "C", "D", "E"),
    col = heat.colors(5), main = "Proportions")

# Sidebar plots (8-10)
# Plot 8: ACF
acf(rnorm(100), main = "ACF")

# Plot 9: QQ plot
qqnorm(rnorm(100), main = "Q-Q Plot", pch = 19, col = "purple")
qqline(rnorm(100), col = "red", lwd = 2)

# Plot 10: Dot chart
dotchart(sort(setNames(sample(10:30, 5), LETTERS[1:5])),
         pch = 19, color = "darkblue", main = "Dot Chart")

customLayout: PowerPoint Integration

customLayout can export layouts directly to PowerPoint slides via the officer package.

library(officer)

# Create layout for PowerPoint
lay <- lay_new(matrix(1:2, nrow = 1))
olay <- phl_layout(lay)

# Create PowerPoint and add plots
pptx <- read_pptx()
pptx <- add_slide(pptx, layout = "Blank")
phl_with_gg(pptx, olay, 1, p1)
phl_with_gg(pptx, olay, 2, p2)
print(pptx, target = "my_presentation.pptx")

customLayout: Layout Splitting

Split an existing field into a sub-layout while preserving aspect ratios.

# Create initial layout
main <- lay_new(matrix(c(1, 1, 2, 3), nrow = 2))
lay_show(main)

# Split field 1 into two parts
sub <- lay_new(matrix(1:2, ncol = 2))
split_layout <- lay_split_field(main, sub, field = 1)
lay_show(split_layout)

patchwork: Annotation and Tagging

patchwork excels at adding panel labels and collecting legends.

(p1 + p2) / (p3 + p4) +
  plot_annotation(
    title = "Combined Figure",
    subtitle = "Four ggplot2 plots arranged in a grid",
    tag_levels = "A"
  )

patchwork: Inset Plots

Easily add inset plots within other plots.

p1 + inset_element(p2, left = 0.6, bottom = 0.6, right = 1, top = 1)

cowplot: Aligned Axes

cowplot provides precise control over axis alignment.

# Align axes across different plot types
plot_grid(p1, p2, p3, p4,
          align = "hv",  # Align both horizontally and vertically
          axis = "tblr")  # Align all axes

cowplot: Drawing Labels

Add custom labels and annotations.

main <- plot_grid(p1, p2, p3, p4,
                  labels = c("A", "B", "C", "D"),
                  label_size = 14)
main

When to Use Each Package

Choose customLayout when:

You need to work with base graphics (not just ggplot2)
You want to export layouts to PowerPoint slides
You prefer matrix-based layout specification similar to base layout()
You need to preserve aspect ratios when combining complex layouts
You’re building layouts programmatically from matrices

Choose patchwork when:

You work exclusively with ggplot2 plots
You prefer intuitive operator syntax (+, /, |)
You need automatic panel labeling (A, B, C, etc.)
You want to collect legends into a single location
You need inset plots within other plots

Choose cowplot when:

You need publication-ready figure composition
You require precise axis alignment across plots
You want fine-grained control over plot positioning
You need consistent theming with the cowplot theme
You’re working on scientific manuscripts with specific formatting

Combining Packages

These packages are not mutually exclusive. You can combine them:

# Use patchwork to create a combined plot
combined_patchwork <- (p1 | p2) / (p3 | p4)

# Use cowplot to add it to a grid with labels
plot_grid(combined_patchwork,
          labels = "AUTO",
          label_size = 12)

# Use cowplot's get_legend with customLayout
library(cowplot)

# Create plots with legends
p_legend <- ggplot(df, aes(x, y, color = group)) +
  geom_point() +
  theme(legend.position = "bottom")

# Extract legend using cowplot
legend <- get_legend(p_legend)

# Create layout with customLayout
lay <- lay_bind_row(
  lay_new(matrix(1:2, nrow = 1)),
  lay_new(1),
  heights = c(4, 1)
)

# Combine plots and legend
plots_no_legend <- list(
  p_legend + theme(legend.position = "none"),
  p4,
  legend
)
lay_grid(plots_no_legend, lay)

Conclusion

Each package offers a unique approach to combining plots in R:

customLayout provides flexibility across graphics systems and PowerPoint integration
patchwork offers the most intuitive syntax for ggplot2 users
cowplot excels at publication-ready figures with precise alignment

Your choice depends on your specific needs: graphics system, output format, and preferred syntax style.

Zygmunt Zawadzki

2026-01-17

Introduction

Feature Comparison Table

Package Overview

customLayout

patchwork

cowplot

Basic Comparison

Using customLayout

Using patchwork

Using cowplot

Complex Layouts

Different Width Columns

customLayout

patchwork

cowplot

Nested Layouts

customLayout

patchwork

cowplot

Unique Features

customLayout: Base Graphics Support

customLayout: PowerPoint Integration

customLayout: Layout Splitting

patchwork: Annotation and Tagging

patchwork: Inset Plots

cowplot: Aligned Axes

cowplot: Drawing Labels

When to Use Each Package

Choose customLayout when:

Choose patchwork when:

Choose cowplot when:

Combining Packages

Conclusion