Keeping two plots vertically aligned with ggplot2/gridExtra
I am trying to merge multiple ggplot2 plots into a single picture, while retaining control over the respective height of each element. The following code works well and it generates a barplot and a curve plot that are perfectly aligned to each other (see here):
library(ggplot2)
library(grid)
library(gridExtra)
# Save plot as PDF (set to NULL initially)
pdf(file = NULL)
# Data frame for the first plot
df_1 <- data.frame(
data = c(0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.2, 0.2, 0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.4, 0.2,
0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.5, 0.1, 0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.2, 0.2,
0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.4, 0.2, 0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.5, 0.1,
0.5, 0.1, 0.15, 0.5, 0.5, 0.4),
base = 1:54,
value = c(0.9557, 0.4407, 0.141, 3.4229, 3.5197, 4.7386, 6.9897, 0.3958, 4.7268, 1.8085,
10.5312, 5.9011, 7.9393, 5.7363, 13.8539, 11.2096, 14.8407, 11.4008, 0.5412, 10.3042,
11.3208, 15.6669, 8.2177, 12.6917, 20.7124, 10.9896, 24.7246, 7.0544, 21.0489, 16.4536,
7.4753, 28.2106, 32.7284, 29.7557, 23.3567, 29.9186, 30.3813, 15.1307, 20.2323, 34.4262,
20.8036, 18.7994, 41.7036, 26.828, 10.7628, 28.2335, 4.4982, 3.6528, 33.8571, 11.6208,
43.54, 26.7859, 22.2176, 11.0588)
)
# First plot with ggplot
plot_1 <- ggplot(df_1, aes(x = base, y = data, fill = value)) +
geom_bar(stat = 'identity', position = position_dodge(width = 0.9)) +
scale_fill_distiller(palette = 'YlGnBu') +
theme(
axis.text.x = element_text(size = 9, angle = 0, vjust = 0.5),
axis.text.y = element_text(size = 9),
legend.position = 'right',
legend.spacing.x = unit(5, 'pt'),
legend.text = element_text(margin = margin(0, 0, 0, 0, 'pt'), size = 10),
legend.key = element_rect(fill = NA),
axis.title.x = element_blank(),
axis.title.y = element_blank(),plot.margin = margin(0,0,0,0)
) +
guides(
colour = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
shape = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
fill = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5)
)+
xlim(0, 55)
# Data frame for the second plot
df_2 <- data.frame(
x1 = c(4, 12, 1, 2),
x2 = c(9, 15, 24, 54),
y1 = c(0, 0, 0, 0),
y2 = c(0, 0, 0, 0),
priority = c('high', 'low', 'mid', 'high')
)
# Second plot with ggplot
plot_2 <- ggplot(df_2) +
geom_curve(aes(x = x1, xend = x2, y = y1, yend = y2, color = priority),
curvature = 1, linewidth = 0.5, ncp = 1000) +
expand_limits(y = -1) +
coord_fixed(ratio = 22, ylim = c(-1, 0)) +
scale_colour_brewer(palette = 'YlGnBu') +
theme(
legend.position = 'right',
legend.spacing.x = unit(5, 'pt'),
legend.text = element_text(margin = margin(0, 0, 0, 0, 'pt'), size = 10),
legend.key = element_rect(fill = NA),
axis.title.x = element_blank(),
axis.title.y = element_blank(),plot.margin = margin(0,0,0,0)
) +
guides(
colour = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
shape = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
fill = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5)
) +
xlim(0, 55)
# List of ggplot objects
plots <- list(plot_1, plot_2) # Add as many plots as you need
# Apply a small margin to each plot to reduce extra white space
plots_adjusted <- lapply(plots, "+", theme(plot.margin = margin(0, 0, 0, 0)))
# Convert adjusted plots to gtables
grob_plots <- lapply(plots_adjusted, ggplotGrob)
# Calculate the maximum widths across all plots
max_widths <- do.call(unit.pmax, lapply(grob_plots, function(g) g$widths))
# Set each plot's widths to the calculated maximum widths
grob_plots <- lapply(grob_plots, function(g) { g$widths <- max_widths; g })
# Arrange the grobs with specific heights
combined_plot <- arrangeGrob(grobs = grob_plots, heights = unit(c(1,1), "null"), padding = unit(0, "pt"))
dev.off()
# Save the combined plot to a file
ggsave("~/Downloads/mytest.pdf", combined_plot,dpi = 300)
However, when I try to modify the width on the output PDF file in ggsave() (for example by setting width=20), the relative proportions of the plots are lost, and only the first one is extended to fill the print area (see here).
Can you advise me on how to fix this?
Solution
As pointed out in the comments, combining the plots using patchwork will keep them aligned. To control the height of each plot, you can use the height argument in the patchwork::plot_layout function. However, this only works if you remove the coord_fixed argument from plot_2.
library(ggplot2)
library(patchwork)
# Data frame for the first plot
df_1 <- data.frame(
data = c(0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.2, 0.2, 0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.4, 0.2,
0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.5, 0.1, 0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.2, 0.2,
0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.4, 0.2, 0.25, 0.5, 0.1, 0.15, 0.5, 0.5, 0.5, 0.1,
0.5, 0.1, 0.15, 0.5, 0.5, 0.4),
base = 1:54,
value = c(0.9557, 0.4407, 0.141, 3.4229, 3.5197, 4.7386, 6.9897, 0.3958, 4.7268, 1.8085,
10.5312, 5.9011, 7.9393, 5.7363, 13.8539, 11.2096, 14.8407, 11.4008, 0.5412, 10.3042,
11.3208, 15.6669, 8.2177, 12.6917, 20.7124, 10.9896, 24.7246, 7.0544, 21.0489, 16.4536,
7.4753, 28.2106, 32.7284, 29.7557, 23.3567, 29.9186, 30.3813, 15.1307, 20.2323, 34.4262,
20.8036, 18.7994, 41.7036, 26.828, 10.7628, 28.2335, 4.4982, 3.6528, 33.8571, 11.6208,
43.54, 26.7859, 22.2176, 11.0588)
)
# First plot with ggplot
plot_1 <- ggplot(df_1, aes(x = base, y = data, fill = value)) +
geom_bar(stat = 'identity', position = position_dodge(width = 0.9)) +
scale_fill_distiller(palette = 'YlGnBu') +
theme(
axis.text.x = element_text(size = 9, angle = 0, vjust = 0.5),
axis.text.y = element_text(size = 9),
legend.position = 'right',
legend.spacing.x = unit(5, 'pt'),
legend.text = element_text(margin = margin(0, 0, 0, 0, 'pt'), size = 10),
legend.key = element_rect(fill = NA),
axis.title.x = element_blank(),
axis.title.y = element_blank(),plot.margin = margin(0,0,0,0)
) +
guides(
colour = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
shape = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
fill = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5)
)+
xlim(0, 55)
# Data frame for the second plot
df_2 <- data.frame(
x1 = c(4, 12, 1, 2),
x2 = c(9, 15, 24, 54),
y1 = c(0, 0, 0, 0),
y2 = c(0, 0, 0, 0),
priority = c('high', 'low', 'mid', 'high')
)
# Second plot with ggplot
plot_2 <- ggplot(df_2) +
geom_curve(aes(x = x1, xend = x2, y = y1, yend = y2, color = priority),
curvature = 1, linewidth = 0.5, ncp = 1000) +
expand_limits(y = -1) +
coord_fixed(ratio = 22, ylim = c(-1, 0)) +
scale_colour_brewer(palette = 'YlGnBu') +
theme(
legend.position = 'right',
legend.spacing.x = unit(5, 'pt'),
legend.text = element_text(margin = margin(0, 0, 0, 0, 'pt'), size = 10),
legend.key = element_rect(fill = NA),
axis.title.x = element_blank(),
axis.title.y = element_blank(),plot.margin = margin(0,0,0,0)
) +
guides(
colour = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
shape = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5),
fill = guide_legend(label.position = 'right', title.position = 'top', title.hjust = 0.5)
) +
xlim(0, 55)
p <- plot_1 / plot_2
# Save the combined plot to a file
ggsave("~/Downloads/mytest.pdf",
p,
width = 20, height = 10)
If you want to adjust the heights, you can do:
p <- plot_1 / plot_2 + plot_layout(heights = c(2,1))
p
To use a large width in ggsave, one option is to change the aspect ratio of the plot:
p <- plot_1 / plot_2 &
theme(aspect.ratio = 1/10)
ggsave("~/Downloads/mytest.pdf",
p,
width = 40, height = 10)
- How to use purrr:map() and rlang to emulate a pipe chain
- magick annotate picture with arrows
- Import png files and convert to animation(.mp4) in R
- Convert a column with text files into separate images in R
- ggplot2 x-axis with many hours for each of many days. Is there a way to span the dates in the x-axis over the hours for that day?
- R odbc::odbcListDrivers() does not list dirver in /opt/homebrew/etc/odbcinst.ini
- Change size and aspect ratio without distortion
- How can I get an R environment via Sys.getenv() with GitHub Actions using secrets?
- ggplot geom_point color based on both x and y axis values
- Perform a random binomial draw for each row in R without rowwise()
- Scrape the university name (in QS World University Rankings website) with R
- ggplot for linear-log regression model?
- How to join (merge) data frames (inner, outer, left, right)
- Avoid rescaling while binning using scale_*_steps
- How can I mock a function globally using testthat?
- R - could not find function "cld"
- Create multiple lagged variables with different offsets
- Expanding dataframe to include non existing values
- Split string to columns based on paragraph ending from ocr'd image
- from magick-image to rasterBrick
- How to remove repeated elements in a vector, similar to 'set' in Python
- Rename multiple variables at once using dplyr
- Calculate average distance to coastline in R
- Reading large multi-part table from file and combing its parts into one tibble
- Processing multiple images with Magick (in R) with transformations
- R: Convert/Read 3D Matrix into a 'magick' object and vice versa
- Error using magick R to import PDF
- Method in R to crop whitespace on svg file
- Read table from a website into R Studio and create a dataframe with the info
- Perspective transformation using R and magick