Apply function to several variables by group in R

Question

I have some data that consists in 3 response variables divided in 3 groups tested in several batches . e.g:

Batch    Type    Replicate    Y1     Y2    Y3
1         A        a           200    100   80

I need to extract the variances for that I build a function:

My_function <- function(x) {
  md_mm <- lmer(Y1 ~ (1|Batch), data = x) 
  tab_mm <- summary(md_mm)
  tab2_mm <- as.data.frame(tab_mm$varcor)
  m <- mean(x$Y1,na.rm=TRUE)
  BatchVar <-tab2_mm[1,4]
  ReptVar <- tab2_mm[2,4]
  df_mm <- data.frame(m,BatchVar, ReptVar)
  return(df_mm)
}

and then I use it:

    table <- df %>% 
  group_by(Type) %>% 
  do(My_function(.)) %>% 
  as.data.frame()

So far this works, but I have been trying to change it so I can get it so the table includes the results for all the variables. I tried with a function with My_function(x,y) and then map() so it runs on all the variables and it works if I tried with the whole variable, when I try to group it then if goes wrong. If anyone has an idea that could help, it would be very much appreciated. Thanks,

Hi user51962, welcome to Stack Overflow. It will be much easier for us to help you if you provide a generous sample of `df` with `dput(df)` or if your data is very large `dput(df[1:30,])`. You can edit your question and paste the output. You can surround it with three backticks (```) for better formatting. See [How to make a reproducible example](https://stackoverflow.com/questions/5963269/) for more info. — Ian Campbell, May 13 '20 at 14:00

score 0 · Answer 1 · answered May 13 '20 at 14:08

I created a data, trying to match the one you posted:

library(tidyverse)

set.seed(123)
df <- data.frame(
  batch =  sample(c(1, 2), 100, replace = TRUE), 
  type = sample(c("A", "B"), 100, replace = TRUE), 
  replicate = sample(c("A", "B"), 100, replace = TRUE), 
  y1 = rnorm(100), 
  y2 = rnorm(100),
  y3 = rnorm(100))

Now, you can use group_by and summarise_if to calculate by group only the numeric variables (you can replace var(x) with your actual function):

df %>% 
  group_by(batch, type, replicate) %>%
  summarise_if(is.numeric, function(x) {var(x)})

Results:

 batch type  replicate    y1    y2    y3
  <dbl> <fct> <fct>     <dbl> <dbl> <dbl>
1     1 A     A         0.766 1.68  2.02 
2     1 A     B         0.908 0.917 1.29 
3     1 B     A         0.979 1.53  0.567
4     1 B     B         0.911 0.727 1.33 
5     2 A     A         0.681 1.16  0.871
6     2 A     B         0.627 0.727 0.781
7     2 B     A         0.688 0.505 1.16 
8     2 B     B         1.98  0.890 0.890

Thanks for the reply, I think summarize only works for predefined functions like the one in your example. Mine was made so it doesn't work. — user51962, May 15 '20 at 11:55
You're welcome, however, I think it's not correct. The `function(x)` argument can take any functions. You can find here more info https://dplyr.tidyverse.org/reference/summarise_all.html — DJV, May 16 '20 at 20:57

Apply function to several variables by group in R

1 Answers1