geom_rect with factor variable

[fergusjchadwick]

Edit: Added copy of plot generated
Hey,

Big fan of RStudio and ggplot, apologies for the clunky code below but hopefully it's reproducible.

I'm trying to plot credible intervals for a set of parameters and geom_rect seems to be the neatest way to plot them but means I have to treat the factors as integers but as soon as the factors are out of order things become messy.

Essentially, I want to be able to treat the x axis as a factor without doing the messy integer transformation back and forth.

An option to treat the x variable like geom_tile and the y variable like geom_rect seems to be ideal but I wondered if there's any alternative solution that's currently available.

library(ggplot2)
set.seed(1859)
dat <- data.frame("CredInt" = factor(rep(c("80perc", "60perc", "40perc"), 8)),
                  "Lower_CI" = c(runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
                                 runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
                                 runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
                                 runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
                                 runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
                                 runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
                                 runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
                                 runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4)),
                  "Upper_CI" = c(runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
                                 runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
                                 runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
                                 runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
                                 runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
                                 runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
                                 runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
                                 runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6)), 
                  "Species" = factor(c(rep("A", 12), rep("B", 12))),
                  "Parameter" = factor(paste0("P", c(rep(1,3), rep(2,3), rep(5,3), 
                                                     rep(3,3), rep(4,3), rep(6,3), 
                                                     rep(7,3), rep(8,3)))))




ggplot(dat, aes(x = as.integer(Parameter))) +
  geom_rect(aes(xmin = as.integer(Parameter) - 0.5,
                xmax = as.integer(Parameter) + 0.5,
                ymin = Lower_CI,
                ymax = Upper_CI,
            fill = as.integer(CredInt))) +
  scale_x_continuous(breaks = 1:length(levels(dat$Parameter)),labels = levels(dat$Parameter)) +
  facet_wrap(~Species, scales = "free_x")
 

[technocrat]

I'm a bit too foggy to decide if this is equivalent

suppressPackageStartupMessages({
  library(ggplot2)
})

set.seed(1859)

dat <- data.frame(
  "CredInt" = factor(rep(c("80perc", "60perc", "40perc"), 8)),
  "Lower_CI" = c(
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4),
    runif(1, -10, -8), runif(1, -8, -6), runif(1, -6, -4)
  ),
  "Upper_CI" = c(
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6),
    runif(1, 8, 10), runif(1, 6, 8), runif(1, 4, 6)
  ),
  Group = rep(c("A", "B"), each = 2),
  Parameter = rep(1:8, each = 3))


ggplot(dat, aes(x = Parameter)) +
  geom_rect(aes(xmin = Parameter - 0.5,
                xmax = Parameter + 0.5,
                ymin = Lower_CI,
                ymax = Upper_CI,
                fill = CredInt)) +
  facet_wrap(~ Group, scales = "free_x") + 
  theme_minimal()

[fergusjchadwick]

Thanks for your interest. My copy of the graph produced didn't load for some reason so I've re-added it there. I think your plot is slightly different because the grouping is different, so it should be

Group = rep(c("A", "B"), each = 12)

I think. Thanks for pointing out the 'each' syntax, much neater.