In csdaw/ggvd: Another Implementation of Venn Diagrams in 'ggplot2'

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.width = 5,
  fig.height = (5*5.5)/6,
  out.width = "90%",
  dpi = 150
)

library(ggplot2)
library(ggvd)

This package aims to simplify making 2-4 way Venn diagrams with ggplot2, whilst allowing for easy customisation. It builds upon the ggvenn and ggVennDiagram packages but is (in my opinion) simpler use. In future I hope to adapt the 2-3 way Euler diagrams from VennDiagram as well.

For comparing 5 or more sets, it is better to use UpSet plots (see ggupset, UpSetR or ComplexUpset).

`prepare_venn()`

Starting with a list of numeric or characters vectors (i.e. a list of sets) to compare, you must construct a data.frame of the appropriate format using prepare_venn(). It is important that within each vector in the list, there are no duplicated elements.

# Create dummy data
set.seed(42)
lst <- list(
  "Set 1" = sample.int(2000, size = 1651, replace = FALSE),
  "Set 2" = sample.int(2000, size = 687, replace = FALSE),
  "Set 3" = sample.int(2000, size = 872, replace = FALSE),
  "Set 4" = sample.int(2000, size = 712, replace = FALSE)
)

# Ensure no duplicates within each set, 
# otherwise prepare_venn() will produce an error
sapply(lst, function(x) any(duplicated(x)))

# Create a data.frame to input into ggplot()
venn4_df <- prepare_venn(lst)
venn4_df

# Add a column for discrete fill colours
venn4_df <- prepare_venn(lst, fill = c("blue", "yellow", "green", "red"))
venn4_df

# Create data.frames for 2 and 3-way Venns as well
venn2_df <- prepare_venn(lst[1:2])
venn3_df <- prepare_venn(lst[1:3])

`geom_venn()`

# The most basic outputs
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements), data = venn2_df)
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements), data = venn3_df)
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements), data = venn4_df)

# Use theme_void() to get a plain background
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements), data = venn4_df) + 
  theme_void()

Venn types

# Give circles discrete fills (discrete Venn type)
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements, fill = fill), 
            type = "discrete", data = venn4_df) + 
  scale_fill_identity() + 
  theme_void()

# Fill each overlap based on the count (continuous Venn type)
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements, fill = count),
            type = "continuous", data = venn4_df) + 
  scale_fill_gradientn(colours = alpha(c("white", "red"), 0.7)) + 
  theme_void()

Set names

# Custom set name position
new_pos <- data.frame(
  x = c(-1.5, -1, 1, 1.5),
  y = c(-1.2, 1.5, 1.5, -1.2)
)

ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements, fill = fill), 
            type = "discrete", data = venn4_df, 
            set_name_pos = new_pos) + 
  scale_fill_identity()

Set totals

# Add set totals
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements, fill = fill), 
            type = "discrete", data = venn4_df, 
            set_total = TRUE) + 
  scale_fill_identity() + 
  theme_void()

Segment counts

Segment percentages

# Adjust percentage y position down a little
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements, fill = fill), 
            type = "discrete", data = venn4_df, 
            percentage_nudge = -0.09) + 
  scale_fill_identity() + 
  theme_void()

# Adjust set total y position above
ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements, fill = fill), 
            type = "discrete", data = venn4_df, 
            set_total = TRUE, set_total_pos = c(0, 0.2)) + 
  scale_fill_identity() + 
  theme_void()

Adding annotations

p <- ggplot() + 
  geom_venn(aes(set_name = set_name, elements = elements), 
            type = "discrete", data = venn2_df,
            percentage_nudge = -0.08) + 
  annotate("curve", x = -1.2, xend = -0.1, y = -1.2, yend = -0.2,
           curvature = 0.3, arrow = arrow(length = unit(2, "mm"))) +
  annotate("text", x = -1.25, y = -1.2, label = "Interesting!", hjust = "right") + 
  labs(title = "Venn Diagram")
p
p + theme_void()