I am wondering if it is somehow possible to access the columns of the provided data within a ggplot2
graph for the title. So something like that:
gg
Here's two ways I've done this using split
. You can use split
to split your dataframe into a named list of dataframes, based on a variable. So calling split(mpg, .$manufacturer)
gives you a list of dataframes, where each dataframe is associated with a manufacturer, e.g. split_df$audi
is the dataframe of all observations made by Audi.
library(dplyr)
library(purrr)
library(ggplot2)
split_df <- split(mpg, .$manufacturer)
First way you could do this is to just call ggplot
on a single item in the list. Since the list is named, names(split_df)[1]
will give you the name, "audi".
ggplot(split_df[[1]], aes(x = hwy, y = displ, label = model)) +
geom_point() +
geom_text(data = . %>% filter(hwy > 28)) +
ggtitle(names(split_df)[1])
That's kinda cumbersome, especially if you want plots for multiple manufacturers. When I've done this, I've used the map functions from purrr
. What's really cool is imap
, which maps over both the list and its names. Here I'm making a list of plots by mapping over the list of dataframes; each plot gets a title from the name of that list item.
plots <- imap(split_df, function(df, manufacturer) {
ggplot(df, aes(x = hwy, y = displ, label = model)) +
geom_point() +
geom_text(data = . %>% filter(hwy > 28)) +
ggtitle(manufacturer)
})
plots$audi
Then I can pull up a specific item from that list of plots. This is also handy for if you need to use walk
to map over them and save every plot or a subset of plots, or if you need to use a grid
function to arrange them into output, or really anything else cool that purrr
is great for.