Creating a 5-way table in R

I’m hoping to create a 5-way table that displays counts of overlap between 5 binary variables.

Something like a correlation table matrix, but with counts of the number of times both of those binary variables were = 1.

So the table would have column headers a, b, c, d, e and rows a, b, c, d, and e.

I was thinking of creating a matrix using count variables, but am open to other suggestions.

Thank you!

I’ve tried using various table functions, but am struggling with only getting counts to display

If the columns in a data frame are binary, one can create a table of counts where two columns both have the value of one as follows.

# set seed for reproducibility
set.seed(91701987)

# create a data frame of 5 columna

df <- data.frame(a = runif(100),
                 b = runif(100),
                 c = runif(100),
                 d = runif(100),
                 e = runif(100))

# convert to 0 / 1 values
df$a <- ifelse(df$a > .3,1,0)
df$b <- ifelse(df$b > .4,1,0)
df$c <- ifelse(df$c > .5,1,0)
df$d <- ifelse(df$d > .6,1,0)
df$e <- ifelse(df$e > .7,1,0)

Now that we've created some data, we'll use matrix multiplication to find the counts where pairs of columns both have the value of 1.

as.matrix(t(df)) %*% as.matrix(df)

...and the output:

> as.matrix(t(df)) %*% as.matrix(df)
   a  b  c  d  e
a 67 40 32 23 26
b 40 61 27 22 19
c 32 27 48 18 18
d 23 22 18 36 12
e 26 19 18 12 35
> 

Now, we'll check the A / B cells (2,1) and (1,2):

# check the a / b value
t(df$a) %*% df$b

...and the output, which matches the cells in the matrix above:

> t(df$a) %*% df$b
     [,1]
[1,]   40