combination values in R for venn diagram

Question

I want to plot circles within circles. I have been successful creating a triple with the following breakdown in my code:

venn.plot <- draw.triple.venn(
area1 = 1171,
area2 = 979,
area3 = 579,
n12 = 979,
n13 = 579,
n23 = 579,
n123 = 579,

The smallest circle is 579, the next circle adds 400 to be a total of 979 and the next adds 192 to be a total of 1171.

I want to extend this to 5 levels, with the next adding 52 to be 1223 and the last adding 11 to be 1234.

I can not work out the combination values to successfully produce a diagram. The error I get is

ERROR [2018-05-04 16:05:48] Impossible: a5 <- area5 - a6 - a7 - a15 - a16 - a17 - a18 - a25 - a26 - a27 - a28 - a31 - a20 - a29 - a21 - a10 produces negative area Error in draw.quintuple.venn(area1 = 1234, area2 = 1223, area3 = 1171,  :   Impossible: a5 <- area5 - a6 - a7 - a15 - a16 - a17 - a18 - a25 - a26 - a27 - a28 - a31 - a20 - a29 - a21 - a10 produces negative area

Obviously I am adding up wrong.

What are the n values I should be using for these? n12 = n13 = n14 = n15 = n23 = n24 = n25 = n34 = n35 = n45 = n123 = n124 = n125 = n134 = n135 = n145 = n234 = n235 = n245 = n345 = n1234 = n1235 = n1245 = n1345 = n2345 = n12345 =

Answer 1

If I understand you would like each set to be a subset of another. Here is the call:

library(VennDiagram)
venn.plot <- draw.quintuple.venn(
  area1 = 579L,
  area2 = 979L,
  area3 = 1171L,
  area4 = 1223L,
  area5 = 1234L,
  n12 = 579L,
  n13 = 579L,
  n14 = 579L,
  n15 = 579L,
  n23 = 979L,
  n24 = 979L,
  n25 = 979L,
  n34 = 1171L,
  n35 = 1171L,
  n45 = 1223L,
  n123 = 579L,
  n124 = 579L,
  n125 = 579L,
  n134 = 579L,
  n135 = 579L,
  n145 = 579L,
  n234 = 979L,
  n235 = 979L,
  n245 = 979L,
  n345 = 1171L,
  n1234 = 579L,
  n1235 = 579L,
  n1245 = 579L,
  n1345 = 579L,
  n2345 = 979L,
  n12345 = 579L)

each area intersection is equal to the minimum of the areas participating in it.

Now this is too cumbersome especially if you need to perform it often, here is a programmatic way:

define your set:

val = c(579, 979, 1171, 1223, 1234)

name it

names(val) ;- LETTERS[1:length(val)]

maybe there is a more elegant way, but from the top of my head:

z <- lapply(1:length(val),function(x){
  comb <- combn(names(val), m = x) #all combination of 1:5 elements of names
  value <- apply(comb, 2, function(i){ #for each combination return minumum from val vector
    return(min(val[names(val) %in% i]))
  })
  set <- apply(comb, 2, paste0, collapse = "&", sep = "") #paste the names with "&" in beetwean
  names(value) = set
  return(value)
})

run each line of code separately (use 3 instead of x for instance) to understand what it does.

z <- unlist(z)

library(eulerr) # a nice library
plot(euler(z, input = "union"), quantities = T)

lets make it into a function:

create_eulerr_data <- function(val){
  z <- lapply(1:length(val),function(x){
  comb <- combn(names(val), m = x)
  value <- apply(comb, 2, function(i){
    return(min(val[names(val) %in% i]))
  })
  set <- apply(comb, 2, paste0, collapse = "&", sep = "")
  names(value) = set
  return(value)
  })
  z <- unlist(z)
  }

now the good thing about this is that it will work for any number of sets:

val <- c(26, 53, 78)
names(val) <- LETTERS[1:length(val)]
plot(euler(create_eulerr_data(val),
           input = "union"),
     quantities = T)

or

val <- c(26, 53, 78, 112, 134, 158)
names(val) <- LETTERS[1:length(val)]
plot(euler(create_eulerr_data(val),
           input = "union"),
     quantities = T)

it will work even if you provide the values unarranged:

val <- c(26, 78, 53)
names(val) <- LETTERS[1:length(val)]
plot(euler(create_eulerr_data(val),
           input = "union"),
     quantities = T)