ggplot : multiple lines starting from the same point

Question

I have a dataset like this one :

df <- data.frame(Species = c("Sp A", "Other sp", "Other sp", "Other sp",
                       "Sp A", "Other sp", "Other sp"), 
           Study = c("A", "A", "A", "A", 
                     "B", "B", "B"), 
           Value = c(1, 3, 4, 5, 3, 6, 7))

Looks like this :

> df
   Species Study Value
1     Sp A     A     1
2 Other sp     A     3
3 Other sp     A     4
4 Other sp     A     5
5     Sp A     B     3
6 Other sp     B     6
7 Other sp     B     7

I would like to compare the values of the species A to the values of the other species from a same study.
This is the plot I have now :

ggplot(df, aes(y = Value, x = Species)) + 
    geom_point(shape = 1) +
    geom_line(aes(group = Study), color = "gray50") +
    theme_bw()

I don't want the vertical lines. Instead of this, I would like to have 5 lines, 3 starting from the lower "Sp A" point toward the 3 corresponding "Other sp" points from the same study (A) and 2 starting from the upper "Sp A" toward the 2 other points from the same study (B).

Answer 1

There may be a more elegant way to do it, but here's a solution that works. The trick is to rearrange the data slightly and then use geom_segment:

library(tidyverse)
df <- data.frame(Species = c("Sp A", "Other sp", "Other sp", "Other sp",
                             "Sp A", "Other sp", "Other sp"), 
                 Study = c("A", "A", "A", "A", 
                           "B", "B", "B"), 
                 Value = c(1, 3, 4, 5, 3, 6, 7))

# Split data
other <- df %>% filter(Species == "Other sp")
sp_a <- df %>% filter(Species == "Sp A")

# Recombine into the data set for plotting
df <- other %>%
  inner_join(sp_a, by = "Study")

# Make the plot
ggplot(df, aes(x = Species.y, y = Value.y)) +
  geom_segment(aes(xend = Species.x, yend = Value.x, colour = Study))

Adjust the plot as desired!

Answer 2

You can do this with geom_line if, for each level of Study, you duplicate the number of Sp A rows to match the number of Other sp rows and then assign a unique ID to each pair of Sp A and Other Sp for the group aesthetic.

library(tidyverse)

df %>% group_by(Study) %>% 
  slice(c(rep(1,length(Species[Species=="Other sp"])),2:n())) %>%  
  group_by(Study, Species) %>% 
  mutate(ID = paste0(Study, 1:n())) %>% 
  ggplot(aes(y = Value, x=Species, group=ID, colour=Study)) + 
    geom_point(shape=1) +
    geom_line() +
    theme_bw()