Make columns from geom_col() the same width across output images

[peperazzi]

We did a survey on e-voting and the expected cost. I used patchwork to plot the resulting column plot, but the width of the columns changes from plot to plot due to different number of answers.

Is there a way to make the columns the same width within the figure, without distorting the titles/questions?

Example: Question 1 has two options, while questions 2 and 3 have 4 and 5 options. I would like the bars to have the same width regardless of # of options.

[FJCC]

I think this method at least comes close. Pick a baseline column width (I picked 0.8 where 1 has adjacent columns touching), divide that by the maximum number of categories in any of your plots and multiply by the number of categories in the given plot.

library(ggplot2)
DF <- data.frame(A = c("Yes", "No"), B = c(80,20))
DF2 <- data.frame(A = c("Yes", "No","maybe","perhaps"), B = c(10,20,30,40))
DF3 <- data.frame(A = c("Yes", "No","maybe","perhaps","heck no"), B = c(10,20,30,40, 33))
ggplot(DF, aes(A, B)) + geom_col(width = 0.8/5*2) #DF has 2 categories

ggplot(DF2, aes(A, B)) + geom_col(width = 0.8/5*4) #DF2 has 4 categories

ggplot(DF3, aes(A, B)) + geom_col(width = 0.8/5*5) #DF3 has 5 categories

^{Created on 2024-02-07 with reprex v2.0.2}

[peperazzi]

That solves the issue of making the columns equal width, but now there is a lot a space between the "Yes" and "No" of the top chart, and increasingly smaller sizes when there are more options. Is there a way to adjust the height of of the panel, based on the number of options?

[dromano]

Hi Rob,

Are you trying to make a report with RMarkdown (or Quarto), or are you trying to embed image files in a document some other way?

[peperazzi]

I'm trying to make an image to export.

[nirgrahamuk]

This seems to work as I would expect

library(tidyverse)

d1 <- data.frame(Q = c("Yes","No"),
                 num = c(5,10))

d2 <- data.frame(Q = letters[1:7],
                 num = 1:7 * 5)
g1 <- ggplot(d1) + aes(x=Q,
                       y=num) + geom_col() + coord_flip()

g2 <- ggplot(d2) + aes(x=Q,
                       y=num) + geom_col() + coord_flip()

 

library(patchwork)

g1 + g2 + plot_layout(ncol = 1,heights = c(length(d1$Q),
                                           length(d2$Q)))

[peperazzi]

That works, and it doesn't even distort the title text.

Thanks.