Mean by groups by multiple columns in r-CodePudding

I have the next dataframe

obs1   obs2   obs3   zone
1       0      1     Rural
1       1      1     Rural
0       1      1     Urban
1       0      0     Urban
0       1      0     Rural

I am trying to get something like this

Mean Rural    0.6666
Mean Urban    0.5

Is that possible?

CodePudding user response：

You can try:

aggregate(values ~ zone, cbind(stack(df[-4]), df[4]), mean)

   zone    values
1 Rural 0.6666667
2 Urban 0.5000000

Or (assuming no missing values):

tapply(rowMeans(df[-4]), df[4], mean)

    Rural     Urban 
0.6666667 0.5000000

CodePudding user response：

To add an {dplyr} approach:

library(dplyr)

dat %>% 
  group_by(zone) %>% 
  summarise(values = mean(c_across(obs1:obs3)))

#> # A tibble: 2 x 2
#>   zone  values
#>   <chr>  <dbl>
#> 1 Rural  0.667
#> 2 Urban  0.5


# data
dat <- tibble(obs1 = c(1,1,0,1,0),
       obs2 = c(0,1,1,0,1),
       obs3 = c(1,1,1,0,0),
       zone = c("Rural", "Rural", "Urban", "Urban", "Rural"))

^{Created on 2022-05-24 by the reprex package (v2.0.1)}