More efficient approach to create a dummy coding - r

Question

Question: In Python, I would work with dictonaries and use lots of map/apply functions. However, with R I started with this simple approach using lists, and I would like to know whether there is a more efficient/more elegant approach to doing the following.
In statistics, you use dummy variables to represent levels of a nominal attribute. E.g., A/B/C would become 00, 01, 10. A/B/C/D would become 000, 001, 010, 100. Thus, only one 1 per item is allowed. You therefore need n-1 numbers to represent n variables/letters.
Here I create some data:
data <- data.frame(
"upper" = c(1,1,1,2,2,2,3,3,3), # var 1
"country" = c(1,2,3,1,2,3,1,2,3), # var 2
"price" = c(1,2,3,2,3,1,3,1,2) # var 3
)
Create a list with keys (attributes) and values (lists of unique attribute levels):
lst <- list()
for (attribute in colnames(data)) {
lst[[attribute]] = unique(data[[attribute]])
}
Create dummy coding, i is used to consider only n-1 items:
dummy <- list()
for (attribute in colnames(data)) {
i <- 1
for (level in lst[[attribute]]) {
if (length(lst[[attribute]])!=i) {
dummy[[paste0(attribute, level)]] <- ifelse(
data[[attribute]]==level,
1,
0
)
}
i <- i + 1
}
}
Results:
dummy
$upper1
[1] 1 1 1 0 0 0 0 0 0
$upper2
[1] 0 0 0 1 1 1 0 0 0
$country1
[1] 1 0 0 1 0 0 1 0 0
$country2
[1] 0 1 0 0 1 0 0 1 0
$price1
[1] 1 0 0 0 0 1 0 1 0
$price2
[1] 0 1 0 1 0 0 0 0 1

Answer 1

We create a design matrix using model.matrix, split the columns to create a list of list, finally, concatenate the list elements together (do.call(c,..).
res <- do.call("c",lapply(data, function(x) {
x1 <- model.matrix(~0+factor(x))
split(x1, col(x1))}))
As we only need the first two levels, we can subset the 'res' using c(TRUE, TRUE, FALSE) which will recycle to the end of the list.
res[c(TRUE, TRUE, FALSE)]
#$upper.1
#[1] 1 1 1 0 0 0 0 0 0
#$upper.2
#[1] 0 0 0 1 1 1 0 0 0
#$country.1
#[1] 1 0 0 1 0 0 1 0 0
#$country.2
#[1] 0 1 0 0 1 0 0 1 0
#$price.1
#[1] 1 0 0 0 0 1 0 1 0
#$price.2
#[1] 0 1 0 1 0 0 0 0 1