How do you add weights to a boxplot with ggplot2? - r

Question

I´m working on a project that involves analysing the income of those who belong to the active population on a certain territorial area (an Autonomous Community). I need to create a histogram with the sample weights given to me and ggplot2. However,when I try to implement the argument "weight" to the aesthetics, it doesn´t work because whether I include the argument "weight" or not, it plots the same graph. Apart from that, I don´t know how to add the weighted mean since my graph doesn´t even take into account the weights of the sample
This is the code to generate all the data from the territorial area:
rm(list=ls(all=TRUE))
if (!require(sae)) install.packages("sae")
library(sae)
data(incomedata)
help("incomedata")
set.seed(100452840)
cual = sample(1:17,1)
(cual)
datosECV=incomedata
datosECVmas16 = subset(datosECV, (datosECV$labor>0))
datosECVmas16$age = datosECVmas16$age - 1
nrows = dim(datosECVmas16)[[1]]
datosECVmas16$horas = round(rnorm(nrows,34,3), 1)
datosECVmas16$horas[(datosECVmas16$labor==2) | (datosECVmas16$labor == 3)] = 0
datosECVmas16$income = round(jitter(datosECVmas16$income),1)
datosECVmas16$income[datosECVmas16$labor==2] = datosECVmas16$income[datosECVmas16$labor==2]*0.7
datosECVmas16$income[datosECVmas16$labor==3] = 0
datosFinal =
data.frame(ca=datosECVmas16$ac, prov=datosECVmas16$prov,
provlab=datosECVmas16$provlab, gen=datosECVmas16$gen,
edad=datosECVmas16$age, nac=datosECVmas16$nat,
neduc=datosECVmas16$educ, sitemp=datosECVmas16$labor,
ingnorm=datosECVmas16$income, horas=datosECVmas16$horas,
factorel=round(datosECVmas16$weight,1))
datos_ComValenciana = datosFinal[datosFinal[,1]==10,]
if(cual == 1) {
write.table(datos_Andalucia,"datos_Andalucia.txt",row.names=FALSE)
} else if(cual == 2) {
write.table(datos_Aragon,"datos_Aragon.txt",row.names=FALSE)
} else if(cual == 3) {
write.table(datos_Asturias,"datos_Asturias.txt",row.names=FALSE)
} else if(cual == 4) {
write.table(datos_Baleares,"datos_Baleares.txt",row.names=FALSE)
} else if(cual == 5) {
write.table(datos_Canarias,"datos_Canarias.txt",row.names=FALSE)
} else if(cual == 6) {
write.table(datos_Cantabria,"datos_Cantabria.txt",row.names=FALSE)
} else if(cual == 7) {
write.table(datos_CastillaLeon,"datos_CastillaLeon.txt",row.names=FALSE)
} else if(cual == 8) {
write.table(datos_CastillaLaMancha,"datos_CastillaLaMancha.txt",row.names=FALSE)
} else if(cual == 9) {
write.table(datos_Catalunya,"datos_Catalunya.txt",row.names=FALSE)
} else if(cual == 10) {
write.table(datos_ComValenciana,"datos_ComValenciana.txt",row.names=FALSE)
} else if(cual == 11) {
write.table(datos_Extremadura,"datos_Extremadura.txt",row.names=FALSE)
} else if(cual == 12) {
write.table(datos_Galicia,"datos_Galicia.txt",row.names=FALSE)
} else if(cual == 13) {
write.table(datos_ComMadrid,"datos_ComMadrid.txt",row.names=FALSE)
} else if(cual == 14) { write.table(datos_RegMurcia,"datos_RegMurcia.txt",row.names=FALSE)
} else if(cual == 15) {
write.table(datos_ComForalNavarra,"datos_ComForalNavarra.txt",row.names=FALSE)
} else if(cual == 16) {
write.table(datos_PaisVasco,"datos_PaisVasco.txt",row.names=FALSE)
} else {
write.table(datos_Rioja,"datos_Rioja.txt",row.names=FALSE)
}
datosFinal=datosFinal[-12086,]
datos_ComValenciana = datosFinal[datosFinal[,1]==10,]
datosFinal=datosFinal[datosFinal$sitemp<3,]
datos_ComValenciana = datosFinal[datosFinal[,1]==10,]
N <- sum(datos_ComValenciana[,"factorel"])
Important note: the weights are the ones given by the 11th column of the dataset (the dataset in this case is called datos_ComValenciana). This column is the one called "factorel". "Ingnorm" are the different incomes of different people.
This code belowe should get the job done but doesn´t:
ggplot(data = datos_ComValenciana, aes(x = ingnorm,y = ..density..,weight=factorel)) +
geom_histogram(fill="#5DC863FF",alpha=0.6,col="black",bins=18)+
xlab("Ingresos normalizados")+
ylab("Cuenta")+
scale_fill_viridis(alpha=1,discrete=TRUE, option="D")+
ggtitle("Income without the weights")`

Answer 1

You can directly apply any weigths in the aes() function, like this (I made a toy example with mtcars):
library(ggplot2)
library(viridis)
data("mtcars")
mtcars$wt_norm <- mtcars$wt / mean(mtcars$wt)
ggplot(data = mtcars, aes(x = mpg * wt_norm, y = after_stat(density))) +
geom_histogram(fill="#5DC863FF", alpha = 0.6, col = "black", bins = 18)+
xlab("mpg normalizado")+
ylab("Cuenta")+
scale_fill_viridis(alpha = 1, discrete = TRUE, option = "D") +
ggtitle("Consumo normalizado por peso")
Which yields a different result than:
ggplot(data = mtcars, aes(x = mpg, y = after_stat(density))) +
geom_histogram(fill="#5DC863FF", alpha = 0.6, col = "black", bins = 18)+
xlab("mpg")+
ylab("Cuenta")+
scale_fill_viridis(alpha = 1, discrete = TRUE, option = "D") +
ggtitle("Consumo")
I'm by no means an specialist in ggplot, but the argument weights, if it is valid, it seems not to be working.