Standardize variables using dplyr [r]

Question

I would like to standardize variables in R. I know about multiple approahces how this can be done. However, I realy like using this approach bellow:

library(tidyverse)

df <- mtcars

df %>% 
  gather() %>% 
  group_by(key) %>% 
  mutate(value = value - mean(value)) %>% 
  ungroup() %>% 
  pivot_wider(names_from = key, values_from = value)

For some reason this approach does not work since I am not able to return the data to the original format. Therefore, I would like to ask for advice

Answer 1

According to the current documentation, you should be using across-based syntax to perform operations on a desired subset of columns. You can use everything to select all columns or use any other available qualifier. You should only use group_by verb if your desire is to perform operation on groups. group_by is not right choice for selecting variables.

mtcars %>%
    as_tibble() %>%
    mutate(across(where(is.numeric), ~ . - mean(.)))

As for the actual standardisation or any other operation you want to apply to the subset of columns you can use:

.fns Functions to apply to each of the selected columns. Possible values are:

• NULL, to returns the columns untransformed.
• A function, e.g. mean.
• A purrr-style lambda, e.g. ~ mean(.x, na.rm = TRUE)
• A list of functions/lambdas, e.g. list(mean = mean, n_miss = ~ sum(is.na(.x))

So for scale you can do:

mtcars %>%
    as_tibble() %>%
    mutate(across(where(is.numeric), scale))

or with additional arguments

mtcars %>%
    as_tibble() %>%
    mutate(across(where(is.numeric), scale, center = FALSE))

Side notes

As you can see from ?scale documentation, the function returns matrix. In case of the examples above, you will get matrix with one column if this bothers you, you can do:

mtcars %>%
    as_tibble() %>%
    mutate(across(where(is.numeric),  ~ scale(.)[,1]))

Comparison

>> mtcars %>%
...     as_tibble() %>%
...     mutate(across(where(is.numeric),  ~ scale(.)[,1])) %>% 
...     glimpse()
Rows: 32
Columns: 11
$ mpg  <dbl> 0.15088482, 0.15088482, 0.44954345, 0.21725341, -0.23073453, -0.33028740, -0.96078…
$ cyl  <dbl> -0.1049878, -0.1049878, -1.2248578, -0.1049878, 1.0148821, -0.1049878, 1.0148821, …
$ disp <dbl> -0.57061982, -0.57061982, -0.99018209, 0.22009369, 1.04308123, -0.04616698, 1.0430…
$ hp   <dbl> -0.53509284, -0.53509284, -0.78304046, -0.53509284, 0.41294217, 
...
>> 
>> 
>> mtcars %>%
...     as_tibble() %>%
...     mutate(across(where(is.numeric), scale)) %>% 
...     glimpse()
Rows: 32
Columns: 11
$ mpg  <dbl[,1]> <matrix[32 x 1]>
$ cyl  <dbl[,1]> <matrix[32 x 1]>
$ disp <dbl[,1]> <matrix[32 x 1]>
$ hp   <dbl[,1]> <matrix[32 x 1]>
...