Question

I have the following data, and am having some issues with the y axis ordering =

> str(dat)
'data.frame':   3678 obs. of  41 variables:
 $ highest_bid                 : Factor w/ 140 levels "\\N","0","10",..: 2 2 2 2 2 2 2 2 2 2 ...
 $ age_of_oldest_driver        : Factor w/ 75 levels "18","19","20",..: 66 11 33 24 57 3 17 5 32 22 ...

When I plot highest bid on age, I get the desired plot but the age’s are in order of the factor, which is not what I want. The y axis goes from 180, 19, 200, 2300, 25, 230, 250, etc. Because of values like 19, 25, and 2300, it throws the ordering of the y axis off. See below plot.

ggplot(dat, aes(x=factor(age_of_oldest_driver), y=highest_bid)) + 
      stat_summary(fun.y="mean", geom="bar")

Now, I generated some sample data to play around with this problem, but I can’t seem to reproduce it. Here’s what I did. Once again, this works perfectly.

df=data.frame(score=c(400,200,3000,500,751,630,554,630,100,250,330,5100,4100,800),
              age=c(18,18,23,50,19,39,19,23,22,22,40,35,22,16))
str(df)
ggplot(df, aes(x=factor(age), y=factor(score))) + geom_bar()

library(plyr)
library(ggplot2)
ggplot(ddply(df, .(age), mean), aes(x=factor(age), y=factor(score))) + geom_bar()

Any idea on what I’m doing wrong in the initial code that I published.

Thanks!!

Here’s the wrong image. Notice the y-axis.

Answer 1

You can reorder the original factor (age) before ploting it

df$ageord <- factor(df$age, levels = levels(factor(df$age)), ordered = TRUE )

require(ggplot2)
ggplot(df, aes(x=ageord, y = score)) + geom_bar(stat = "identity")

EDIT : if it doesn’t work

df$ageord <- factor(df$age, levels = sort(unique(df$age)), ordered = TRUE )

ggplot(df, aes(x=ageord, y = score)) + geom_bar(stat = "identity")