[R] (no subject)

Jim Lemon jim at bitwrit.com.au
Fri Sep 10 13:03:21 CEST 2010


On 09/10/2010 07:38 PM, Tanvir Khan wrote:
> Hello,
> I'm trying to do bar plot where 'sex' will be the category axis and
> 'occupation' will represent the bars and the clusters will represent
> the mean 'income'.
>
>     sex     occupation   income
> 1  female          j             12
> 2    male          b            34
> 3    male          j             22
> 4  female          j            54
> 5    male          b           33
> 6  female          b           67
> 7    male          j            89
> 8    male          b          65
> 9  female          j          45
> 10   male          j          32
>
> I can do bar plot where sex is the category axis and the clusters
> represent 'occupation'.
> the code is-
>
>> t<- table(data$sex,data$occupation)
>> barplot(f)
>
> and the barplot where the category axis is 'sex' and the cluster
> represent the mean income and median income. The code is -
>> mean=tapply(data$income,data$sex,mean)
>> mean
>    female     male
> 38.66667 46.50000
>> median=tapply(data$income,data$sex,median)
>> median
> female   male
>    22.5   49.5
>> r=rbind(mean,median)
>> r
>                 female     male
> mean    38.66667     46.5
> median 22.50000    49.5
>> par(fg='red',cex=1.2)
>> barplot(r,col=c('green','yellow'),cex.axis=1.2,col.axis='red',ylim=c(0,120)
>
> But how can I make 'occupation'' to nest inside 'sex' and then the
> cluster to represent the mean income?
> For example I am attaching a pdf plot that is produced by SPSS.
>
Hi Tanvir,
I think you may be looking for barNest:

examp<-data.frame(sex=sample(c("Female","Male"),100,TRUE),
  position=sample(c("Clerical","Custodial","Manager"),100,TRUE),
  income=rnorm(100,30000,5000))
# just show the final bars
barNest(income~sex+position,examp,
  col=list("gray",c("pink","lightblue"),c("blue","green","tan")))
# show the entire nest of bars
barNest(income~sex+position,examp,showall=TRUE,
  col=list("gray",c("pink","lightblue"),c("blue","green","tan")))

Jim



More information about the R-help mailing list