Geom_label_repel: How to pull labels toward 4 corners of a scatterplot

I have a set of product scores I want to visualise in a scatter. What I'm trying to do is have the labels pull to the four separate corners of the chart. In reality, I have 30-40 to space which is why I'm using ggrepel for the labels.

`
library(ggplot2)

peer_score <- c(3.5, 3.5, 2, 2 )
mkt_score <- c(3.5, 2, 3.5, 2)
revenue <- sample(1:75, 4, replace = T)
prod_name <- c("Product 1", "Product 2", "Product 3", "Product 4") 

df <- data.frame(prod_name, revenue, peer_score, mkt_score)

head(df)

ggplot(df, aes(peer_score, mkt_score)) + 
  geom_point(color = "blue", alpha=0.5) +
  geom_label_repel(aes(label=prod_name)) +
  geom_vline(xintercept=3, linetype="dashed", alpha = 0.9, colour = "dark grey") +
  geom_hline(yintercept=3, linetype="dashed", alpha = 0.9, colour = "dark grey") +
  xlim(1,5) +
  ylim(1,5)
`

So I would want 'Product 1' to be pulled to the top right, 'Product 2' to the bottom right, etc

I have tried using nudge_x or nudge_y but that sends labels all in the same direction. Any suggestions are much appreciated!

Solution

Not sure whether that will work for your real case. But at least for your example data one option would be to use e.g. one geom_label_repel for each of the four quadrants which would allow to set the nudging and the alignment individually.

library(ggplot2)
library(ggrepel)
nudge <- .1

ggplot(df, aes(peer_score, mkt_score)) +
  geom_point(color = "blue", alpha = 0.5) +
  geom_label_repel(data = subset(df, peer_score < 3 & mkt_score < 3), aes(label = prod_name), nudge_x = -nudge, nudge_y = -nudge, hjust = 1, vjust = 1) +
  geom_label_repel(data = subset(df, peer_score < 3 & !mkt_score < 3), aes(label = prod_name), nudge_x = -nudge, nudge_y = nudge, hjust = 1, vjust = 0) +
  geom_label_repel(data = subset(df, !peer_score < 3 & mkt_score < 3), aes(label = prod_name), nudge_x = nudge, nudge_y = -nudge, hjust = 0, vjust = 1) +
  geom_label_repel(data = subset(df, !peer_score < 3 & !mkt_score < 3), aes(label = prod_name), nudge_x = nudge, nudge_y = nudge, hjust = 0, vjust = 0) +
  geom_vline(xintercept = 3, linetype = "dashed", alpha = 0.9, colour = "dark grey") +
  geom_hline(yintercept = 3, linetype = "dashed", alpha = 0.9, colour = "dark grey") +
  xlim(1, 5) +
  ylim(1, 5)