Cleaner way to plot multiple bar charts of different outcome variables (R)

Question

I am wondering if there is a better way to produce 4 barcharts of different outcome variables arranged in a grid:

This is the code I used:

library(cowplot)

bar1 <- ggplot(data = subset(data, !is.na(MHQ_Heading_Male_Quartile))) +
  geom_bar(mapping = aes(x = MHQ_Heading_Male_Quartile))
bar2 <- ggplot(data = subset(data, !is.na(AHQ_Heading_Male_Quartile))) +
  geom_bar(mapping = aes(x = AHQ_Heading_Male_Quartile))
bar3 <- ggplot(data = subset(data, !is.na(MHQ_Heading_Female_Quartile))) +
  geom_bar(mapping = aes(x = MHQ_Heading_Female_Quartile))
bar4 <- ggplot(data = subset(data, !is.na(AHQ_Heading_Female_Quartile))) +
  geom_bar(mapping = aes(x = AHQ_Heading_Female_Quartile))

plot_grid(bar1, bar2, bar3, bar4, ncol = 2)

However, there is a lot of repeated code- is there some function or way to create the same plot with ggplot2 in fewer lines?

Answer 1

I would convert relevant columns from wide to long (the ones ending in "_Quartile" ) and then use facet_wrap to show the 4 plots in a 2x2 grid with scales = "free" .

Something like this:

data %>%
    gather(key, value, ends_with("Quartile")) %>%
    filter(!is.na(value)) %>%
    ggplot(aes(value)) +
    geom_bar() +
    facet_wrap(~ key, scales = "free", ncol = 2, nrow = 2)

Answer 2

As mentioned you need to make it a long format using dplyr gather (or reshape package) and then facet over this.

`data %>%
  select( MHQ_Heading_Male_Quartile, AHQ_Heading_Male_Quartile, MHQ_Heading_Female_Quartile, AHQ_Heading_Female_Quartile) %>%
  gather("Type", "Range", MHQ_Heading_Male_Quartile:AHQ_Heading_Female_Quartile) %>%
  filter(!is.na(Range)) %>%
  ggplot(aes(x=Range)) + 
  geom_bar() + 
  facet_wrap(~Type, scales="free")`

I'll leave it to you to clean the graphs up but that's the basic premise.

Answer 3

Extract the column names to be shown into nms and then for each one use qplot to create a ggplot object so that bars is a list of such objects. Then run plot_grid on that.

nms <- grep("Quartile", names(data), value = TRUE)
bars <- lapply(nms, function(nm) qplot(na.omit(data[[nm]]), xlab = nm))
do.call("plot_grid", bars)