CODEWITHSUNDEEP
rggplot2
leeneumann
leeneumann

Reputation: 27

pass variable to ggmosaic

I have a data set called panel which has multiple categorical variables in interest.

I would like to write a program that will plot a mosaic plot for every pair using a for loop.
vars is the vector with the following names of the columns: 

vars
 [1] "region"   "urb"      "sex"      "race"    "grade"   "dt01"   "dt02"     "dt03"    
 [9] "dt06"     "dt07"     "exercise" "kq7"

This is my program:

library(ggmosaic)

for (i in 1:12){
  for (j in 1:12){
    a <- vars[i]
    b <- vars[j]
    if (j > i){
      m <- ggplot(data = panel) 
      m <- m + geom_mosaic(aes(x = product(a), fill = b))
    }
    print(m)
  }
}

This code gives me a a plot with one big chunk.

plot

When I write the column names manually like in the following code it works

ggplot(data = panel) +
  geom_mosaic(aes(x = product(region), fill = urb))

I tries aes_string but it didn't help.

Upvotes: 0

Views: 261

Answers (1)

Z.Lin
Z.Lin

Reputation: 29075

Two suggestions for consideration.

1: Generate your pairs & use a single loop to go through them, rather than a double loop that goes through repeated combinations.

vars <- c("region", "urb", "sex", "race", "grade", "dt01",     
          "dt02", "dt03", "dt06", "dt07", "exercise", "kq7")
vars.pair <- combn(vars, 2, simplify = FALSE) # list of all valid pairs

> head(vars.pair)
[[1]]
[1] "region" "urb"   

[[2]]
[1] "region" "sex"   

[[3]]
[1] "region" "race"  

[[4]]
[1] "region" "grade" 

[[5]]
[1] "region" "dt01"  

[[6]]
[1] "region" "dt02"

2: In aes_string(), make sure x's aesthetic mapping evaluates to "product(variable.name)" rather than product("variable.name").

for(i in seq_along(vars.pair)){

  p1 <- vars.pair[[i]][1]
  p2 <- vars.pair[[i]][2]

  m <- ggplot(data = panel) +
    geom_mosaic(aes_string(x = paste0("product(", p1, ")"),
                           fill = p2))
  print(m)

}

Upvotes: 3

Related Questions