3

I strictly need to use the summarise_at to compute a weighted mean, with weights based on the values of another column

    df %>% summarise_at(.vars = vars(FACTOR,tv:`smart tv/console`), 
                  .funs = weighted.mean, w=INVESTMENT, na.rm=TRUE)

It always shows the error: 'INVESTMENT' is not found.

I then tried with:

df %>%summarise_at(.vars = vars(FACTOR,tv:`smart tv/console`), 
               .funs = weighted.mean, w=vars(INVESTMENT), na.rm=TRUE)

But in this case : Evaluation error: 'x' and 'w' must have the same length.

Why is this? Am I doing anything wrong? Do you have hints to solve this issue? Thanks

3nomis
  • 541
  • 7
  • 17

2 Answers2

2

You can specify the weights directly within the weighted.mean() function, within the call to funs() like so:

data.frame(x=rnorm(100), y=rnorm(100), weight=runif(100)) %>%
      summarise_at(vars(x,y), funs(weighted.mean(., w=weight)))
mmk
  • 121
  • 2
0

Reproducible Example of Weighted Mean:

df = data.frame(x=rnorm(6), y=rnorm(6), c=runif(6))
df
           x           y          c
1 -1.7474440 -0.31007446 0.05848928
2 -1.2224065  2.00200840 0.83287364
3 -0.2794127 -1.03566512 0.59140733
4 -0.3022496  0.05759183 0.05800232
5  0.4875809  0.88804664 0.36282380
6 -0.5468437  0.27501804 0.37218218

wt = c(5,5,4,1,3,5)/15 # This is the weight
wt
[1] 0.33333333 0.33333333 0.26666667 0.06666667 0.20000000
[6] 0.33333333


> weighted.mean(df$c,wt)
[1] 0.5992344

Weighted Mean R Documentation:

weighted.mean {stats}   R Documentation
Weighted Arithmetic Mean
Description
Compute a weighted mean.

Usage
weighted.mean(x, w, ...)


Default S3 method:


weighted.mean(x, w, ..., na.rm = FALSE)
Arguments
x

an object containing the values whose weighted mean is to be computed.


w

a numerical vector of weights the same length as x giving the weights to use for elements of x.


... 
arguments to be passed to or from methods.


na.rm

a logical value indicating whether NA values in x should be stripped before the computation proceeds.


Details
This is a generic function and methods can be defined for the first argument x: apart from the default methods there are methods for the date-time classes "POSIXct", "POSIXlt", "difftime" and "Date". The default method will work for any numeric-like object for which [, multiplication, division and sum have suitable methods, including complex vectors.


If w is missing then all elements of x are given the same weight, otherwise the weights coerced to numeric by as.numeric and normalized to sum to one (if possible: if their sum is zero or infinite the value is likely to be NaN).


Missing values in w are not handled specially and so give a missing value as the result. However, zero weights are handled specially and the corresponding x values are omitted from the sum.


Value
For the default method, a length-one numeric vector.


See Also
mean


Examples


GPA from Siegel 1994


wt <- c(5,  5,  4,  1)/15
x <- c(3.7,3.3,3.5,2.8)
xm <- weighted.mean(x, wt)
rubengavidia0x
  • 289
  • 2
  • 15