How to add text outside a ggplot2 plot using npc coordinates in R-CodePudding

I am using a function in R to subset data and plot each of those subsets, and mapping (purrr::map) the function over a list of the subsets. I would like to be able to label the plots with a letter in the top left, in a position that remains constant despite the varying ranges of the axes. In other words, I want to use a normalized coordinate system, but I want to use negative values.

I have tried using ggpp::geom_text_npc(), which functions perfectly within the plot, but any values outside of the 0 to 1 range seem to just "round" to values inside the plot. For example:

library(tidyverse)
library(purrr)
library(ggpp)

datumz<-data.frame(ex = c(1:5, 11:15),
                   why = c(rep(1:5,2)),
                   kat = c(rep("a", 5),rep("b", 5)))
                   
selex<-list(c("a", "A"),
            c("b", "B"))
                  
grafz<-function(a){

  part<-datumz %>% 
    filter(kat==a[1])

  ggplot(part) 
    geom_point(mapping=aes(ex, why)) 

    coord_cartesian(clip='off') 
    ggpp::geom_text_npc(size=8,aes(npcx=-.2, npcy=1.2, label=a[2]))
  
}

purrr::map(selex,grafz)

Places the text consistently across the two plots (which is good), but I can't find npcx or npcy values that would print outside of the plot box.

CodePudding user response：

The patchwork package has some good functionality for figure tags. See below for how to add to 1) a set of combined figures with automatic sequencing, 2) a single figure with a custom tag, or 3) as a drop-in replacement for your ggpp::geom_text_npc line.

library(patchwork)
ggplot(subset(datumz, kat == "a")) 
  geom_point(mapping=aes(ex, why)) 
  coord_cartesian(clip='off')  
ggplot(subset(datumz, kat == "b")) 
  geom_point(mapping=aes(ex, why)) 
  coord_cartesian(clip='off')  
  plot_annotation(tag_levels = "A")

Or if you want to make a single panel at a time:

ggplot(subset(datumz, kat == "b")) 
  geom_point(mapping=aes(ex, why)) 
  coord_cartesian(clip='off')  
  plot_annotation(tag_levels = list("B"))

So a drop-in replacement for your ggpp::geom_text_npc line could be:

plot_annotation(tag_levels = list(a[2]))