Weighted means for groups in r - using aggregate and weighted.mean functions together

user_12857915 0
Viewed 209

I'm trying to calculate the weighted mean for group variable X1, across all numeric variables, here is some example data

set.seed(123)
X1=rep(c("A", "B", "C"), each = 4)
Y1=as.numeric(seq(1,12,by=1))
Y2=sample(1:5,12,TRUE)
Y3=sample(10:20,12,TRUE)
wgt <- abs(rnorm(12)*10)
df <- data.frame(X1,Y1,Y2,Y3,wgt)

This is the code I've been using to calculate regular mean values for X1

aggregate( df[, sapply(df, is.numeric)] , by=list(df$X1) , FUN=mean, na.rm=TRUE)

I want to calculate weight mean, weight variable is wgt. I tried both of these codes and neither work. I've tried numerous different ways and nothing is working.

aggregate( df[, sapply(df, is.numeric)] , by=list(df$X1) , FUN=weighted.mean(x, w=df$wgt), na.rm = TRUE)
aggregate( df[, sapply(df, is.numeric)] , by=list(df$X1) , FUN=weighted.mean, w=df$wgt, na.rm = TRUE)

I'm unable to adapt the weighted.mean function. Can anyone tell me where I'm going wrong? Can this function even be used in this situation? Any help is greatly appreciated. Thanks

r
3 Answers

Here is a way to compute the weighted means with aggregate called by by().

res <- by(df, df$X1, function(DF){
  aggregate(cbind(Y1, Y2, Y3) ~ X1, DF, function(y, w) 
    weighted.mean(y, w = DF[['wgt']], na.rm = TRUE))
})
do.call(rbind, res)
#  X1        Y1       Y2       Y3
#A  A  2.152503 2.633935 18.93457
#B  B  6.677851 3.589251 16.90102
#C  C 10.194695 2.638378 16.70958
user_82454060

You could use outer to apply weighted.mean crosswise.

gr <- c("A", "B", "C"); ys <- c("Y1", "Y2", "Y3")
WF <- Vectorize(function(x, y) with(df[df$X1 %in% x, ], weighted.mean(get(y), wgt)))
res <- `dimnames<-`(outer(gr, ys, WF), list(gr, ys))
res
#          Y1       Y2       Y3
# A  2.152503 2.633935 18.93457
# B  6.677851 3.589251 16.90102
# C 10.194695 2.638378 16.70958
user_65740380

Here's a dplyr solution that returns the same answers as @Rui. As requested this will operate on all variables that are numeric regardless of their column name.

df %>% 
   group_by(X1) %>% 
   summarise(across(where(is.numeric), 
                    ~ weighted.mean(.x, wgt), 
                    .names = "weighted_mean_{.col}"))
user_91900340
Related