R help - Jul 2013 - Some problems with back to back bar plots in ggplot2

Hi, I'm making a simple population pyramid using two back-to-back bar plots
for 18 different age groups with totals for males and females.

Yesterday I achieved a fairly serviceable plot using the following.

library(ggplot2)
library(reshape2)
library(plyr)
#make sample data
df<-data.frame(ag=c(1:18),males=sample(100:200,18),females=sample(100:200,18))
#melted as per original data set
df<-melt(df,id="ag")
df
#my plot
ggplot(data=df,aes(x=ag))+geom_bar(subset=.(variable=="males"),aes(y=value),stat="identity",fill="#330099")+
 
geom_bar(subset=.(variable=="females"),aes(y=-value),stat="identity",fill="#FF9333")+
  scale_y_continuous(breaks=seq(-200,200,50),labels=abs(seq(-200,200,50)))+
  scale_x_continuous(breaks=seq(1,18,1),labels=abs(seq(1,18,1)))+coord_flip()+
  theme_bw()+xlab("age group")+ylab("population")

Today it threw errors. I updated to R studio 0.97.551.  I also made sure I had
the latest R 3.0.1, and reinstalled the packages I was using. This is all
running on Windows 7.

It still threw errors. So I tried specifying all of the variables with their
full names, which got me a plot.

ggplot(data=df,aes(x=df$ag))+geom_bar(subset=.(df$variable=="males"),aes(y=df$value),stat="identity",fill="#330099")+
 
geom_bar(subset=.(df$variable=="females"),aes(y=-df$value),stat="identity",fill="#FF9333")+
  scale_y_continuous(breaks=seq(-200,200,50),labels=abs(seq(-200,200,50)))+
  scale_x_continuous(breaks=seq(1,18,1),labels=abs(seq(1,18,1)))+coord_flip()+
  theme_bw()+xlab("age group")+ylab("population")

But now my y axis (which in my case is horizontal as I flipped the co-ords)
won't scale properly.  From what I recall, it looked fine yesterday.

I'm trying to get back to a properly scaled population pyramid, which apart
from some cosmetics, was what I needed.

Any help appreciated.

GavinR

> Hi, I'm making a simple population pyramid using two 
> back-to-back bar plots for 18 different age groups with 
> totals for males and females.
I can run both variants without error, though both throw the same warning
regarding stacking beein 'not well defined ..'

However, the axis scaling in your second variant appears to work if you add a
limits=(c(-200,200) statement to the y scale, as in

ggplot(data=df,aes(x=df$ag))+geom_bar(subset=.(df$variable=="males"),aes(y=df$value),stat="identity",fill="#330099")+
 
geom_bar(subset=.(df$variable=="females"),aes(y=-df$value),stat="identity",fill="#FF9333")+
  scale_y_continuous(limits=c(-200,200),
breaks=seq(-200,200,50),labels=abs(seq(-200,200,50)))+
  scale_x_continuous(breaks=seq(1,18,1),labels=abs(seq(1,18,1)))+coord_flip()+
  theme_bw()+xlab("age group")+ylab("population")


I suspect that by specifying the bar values separately from the data supplied in
the data statement, you've confused ggplot's automatic scaling
somewhere. But you may have to wait on hadley Wickam's input to find out
exactly why.

S Ellison



*******************************************************************
This email and any attachments are confidential. Any use...{{dropped:8}}

Thanks, I've now got the plot I was after.  No idea what the stacking
warning
is about, but it does not seem to compromise my output.

Gavin.



--
View this message in context:
http://r.789695.n4.nabble.com/Some-problems-with-back-to-back-bar-plots-in-ggplot2-tp4671744p4671765.html
Sent from the R help mailing list archive at Nabble.com.