Histogram and weights

[Lucas11]

Hello, I am stuck with the weights of a survey.
I have a vector with the weights of each observation. In this case how many people each observation represent. In the other hand I have the data of each variable. To see properly all the values I should weight the value of each variable for the weight vector (as simple as a multiplication).
But things get complicated when I want to graphic some results like an histogram or a density plot because I dont know how to specify that after the multiplication (variable * weight) the new observation should represents more than one person in the histogram. Any help?

[Shorthand]

geom_histogram() accepts a weight = aesthethic ... see https://ggplot2.tidyverse.org/reference/geom_histogram.html

If you want more help than this, you'll really have to post some example code.

[Lucas11]

Thanks. Lets supouse these are my data

library(tidyverse)

my_data <- tibble(Var_1 = c(900, 1500, 350, 1200, 750, 100),
               my_weights = c(2.2, 3.1, 8.2, 4.2, 5.3, 6.8))

The correct way to proceed would be use the "raw" data to create the histogram like here?

ggplot(my_data, aes(Var_1, weight = my_weights,
                          ))+
  geom_histogram()

Or should I create a new weighted variable first and then use it in the plot like here:

my_data %>%
  mutate(Var_1_weighted = Var_1 * my_weights) -> my_data

ggplot(my_data, aes(Var_1_weighted, weight = my_weights,
))+
  geom_histogram()

[StatSteph]

You should do the first way. When using weights, we say that observation represents that many cases which is what the weight statement does. Imagine the variable is height, say 2 meters, and the weight is 3. It means the observation represents 3 2-meter people not 1 6-meter person which is what you are doing when you multiply by the weight.