Skip to contents

UpSet and Venn Diagram for Set Intersections

Source: R/plt_upset.R
plt_upset.Rd

Visualise intersections of binary categorical variables using Venn diagrams and UpSet plots. Returns both plots and a data frame with intersection group assignments.

Usage

plt_upset(
  data,
  vars,
  levels = c("Yes", "yes"),
  colors = NULL,
  output = c("both", "venn", "upset", "data", "all")
)

Arguments

data

A data frame.

vars

Character vector of binary variable names to intersect.

levels

Character vector of values indicating membership (e.g. c("Yes", "yes", "1")). Default c("Yes", "yes").

colors

Character vector of 3 colours for c(sets_bar, top_bar, matrix_dots). Default uses pal_lancet.

output

What to return:

"venn"

Venn diagram only (ggplot).

"upset"

UpSet plot only (non-zero intersections).

"both"

(default) Combined Venn + UpSet side by side.

"data"

Data frame with intersect_group column added.

"all"

Named list with all plots and data.

Value

Depends on output: a ggplot, a combined plot, a data frame, or a named list.

Note

Requires ggVennDiagram and optionally aplot (for combined layout). Install with pak::pak(c("ggVennDiagram", "aplot")).

See also

Other plot: PlotButterfly(), PlotButterfly2(), PlotRankCor(), plt_cat(), plt_cohen(), plt_con(), plt_dist(), plt_radar(), plt_sankey()

Examples

df <- data.frame(
  A = sample(c("Yes","No"), 100, TRUE, c(0.3, 0.7)),
  B = sample(c("Yes","No"), 100, TRUE, c(0.4, 0.6)),
  C = sample(c("Yes","No"), 100, TRUE, c(0.5, 0.5)),
  D = sample(c("Yes","No"), 100, TRUE, c(0.2, 0.8))
)

# Venn diagram
plt_upset(df, vars = c("A","B","C"), output = "venn")


# UpSet plot
plt_upset(df, vars = c("A","B","C","D"), output = "upset")


# Combined (default)
plt_upset(df, vars = c("A","B","C"))
#> Warning: Combined layout failed: Only know how to add <ggplot> and/or <grob> objects.
#> Returning plots separately.


# Get data with intersection groups
result <- plt_upset(df, vars = c("A","B","C"), output = "data")
table(result$intersect_group)
#> 
#>     B     C   B/C     A A/B/C   A/C   A/B  None 
#>    17    15    10     9     9     8     6    26 

# Custom membership levels
df2 <- data.frame(
  x = sample(0:1, 100, TRUE),
  y = sample(0:1, 100, TRUE),
  z = sample(0:1, 100, TRUE)
)
plt_upset(df2, vars = c("x","y","z"), levels = c("1"))
#> Warning: Combined layout failed: Only know how to add <ggplot> and/or <grob> objects.
#> Returning plots separately.


# All outputs
res <- plt_upset(df, vars = c("A","B","C","D"), output = "all")
#> Warning: Combined layout failed: Only know how to add <ggplot> and/or <grob> objects.
#> Returning plots separately.
names(res)
#> [1] "venn"     "upset"    "combined" "data"