Sorting Stacked Bar Plot by Factor in dataframe

Question

Amateur R user here. I looked pretty hard online to see if this question has been answered/asked but I have not found a good answer. I can't post an image since I do not have 10 "reputations"

I want to have a stacked bar plot that orders the x variable based on the percent contribution of the ingestion route (in descending order).

Percent<-c(0.4,0.75,0.8, 0.3,0.1,0.6,0.25,0.5)
Inh<-data.frame(ID=c(rep(1,4),rep(2,4)),Age=factor(rep(1:4,2), label=c("0-1 year", "1-2 years", "2-3 years","3-6 years")), Route=factor(rep(1), label="Inhalation"),    Percent=Percent)

Ing<-data.frame(ID=c(rep(1,4),rep(2,4)),Age=factor(rep(1:4,2), label=c("0-1 year", "1-2 years", "2-3 years","3-6 years")), Route=factor(rep(1), label="Ingestion"),     Percent=1-Percent)

df<-data.frame(rbind(Inh,Ing))
ggplot(df,aes(x=ID,y=Percent,fill=Route))+ geom_bar(stat="identity")+ 
facet_wrap(~Age, scales = "free_x") +
ylab("Percent Contribution") +
labs(title = "Route Contribution to Exposure by Age Groups")

But I want it to look like this which I mocked up manually:

Percent<-c(0.1,0.6,0.25, 0.3,0.4,0.75,0.8,0.5)
Inh<-data.frame(ID=c(rep(1,4),rep(2,4)),Age=factor(rep(1:4,2), label=c("0-1 year", "1-2 years", "2-3 years","3-6 years")), Route=factor(rep(1), label="Inhalation"),    Percent=Percent)

Ing<-data.frame(ID=c(rep(1,4),rep(2,4)),Age=factor(rep(1:4,2), label=c("0-1 year", "1-2 years", "2-3 years","3-6 years")), Route=factor(rep(1), label="Ingestion"),     Percent=1-Percent)

df<-data.frame(rbind(Inh,Ing))
ggplot(df,aes(x=ID,y=Percent,fill=Route))+ geom_bar(stat="identity")+ 
facet_wrap(~Age, scales = "free_x") +
ylab("Percent Contribution") +
labs(title = "Route Contribution to Exposure by Age Groups")

Thank you in advance!

UPDATE: Thanks to Roland, I have a plot! Questions remain on clarity though. For those interested here's code and final product:

ggplot(df,aes(x=id2,y=Percent,fill=Route, width=1,order = -as.numeric(Route)))+ 
geom_bar(stat="identity")+ 
facet_wrap(~Age, scales = "free_x") +
xlab(" ")+
ylab("Percent Contribution") +
theme(axis.text.x = element_blank(), axis.ticks.x= element_blank() ) +
labs(title = "DEHP Route Contribution to Exposure by Age Groups")

Answer 1

This changes the order without changing the data (as you do in your mock-up). The idea is to create an ordered (by Percent) factor giving the interaction of Age and ID and use this for plotting, but change the axis labels to match only the ID values.

df <- df[order(df$Route,df$Percent),]
df$id2 <- factor(paste(df$ID,df$Age),levels=unique(paste(df$ID,df$Age)),ordered=TRUE)

ggplot(df,aes(x=id2,y=Percent,fill=Route))+ 
  geom_bar(stat="identity")+ 
  scale_x_discrete(labels = setNames(regmatches(levels(df$id2),regexpr("[[:alnum:]]*",levels(df$id2))),levels(df$id2))) +
  facet_wrap(~Age, scales = "free_x") +
  xlab("ID") +
  ylab("Percent Contribution") +
  labs(title = "Route Contribution to Exposure by Age Groups")

However, I think the resulting plot is confusing and difficult to read.

Answer 2

A fundamental question to understand is whether the order is a property of the graph or a property of the data itself. R tends towards the property of the data instead of the plot, so the plotting functions don't have arguments for reordering the parts since that should be done when creating or editing the data. The reorder function is one way to reorder the levels of a factor to use for future graphs/analyses.