Message-ID: <06a4519c-7948-6931-c72f-e02f6f5ea886@auckland.ac.nz>
Date: 2017-06-18T00:30:25Z
From: Rolf Turner
Subject: [FORGED] About error bars on barplots
In-Reply-To: <CAN5afy8hgLqoBgbRMw5tZptKrD5aXFmh4j2XYMvdB2a6ixcH+g@mail.gmail.com>
On 18/06/17 12:10, lily li wrote:
> Hi R users,
>
> I have a question about adding uncertainty bars to stacked bar plots.
>
> DF:
> year A B C Amin Amax Bmin Bmax Cmin Cmax
> 2009 40 45 15 30 61 23 56 14 17
> 2010 36 41 23 26 54 22 51 22 24
>
> I use the code below:
>
> DF.refm = melt(subset(DF[,c(1:4)]),id.vars='year',variable_name='Legend')
>
> fig1 = ggplot(data=DF.refm, aes(x=year,y=value,fill=Legend))+
> geom_bar(stat='identity',size=.5)+ theme_bw()+ xlab('Year')+
> ylab('Percent (%)')
> print(fig1)
>
> But I don't know how to change it a little bit. For example, how to add the
> error bars on the plot, as is shown in the figure below? The error bar for
> variable A ranges from Amin to Amax, the same applies to variables B and C.
> Thanks for your help.
Your figure did not come through, at least not to me. You are probably
doing something fancy, and the mailing list software stripped out the
figure. *Do* learn to keep things simple. _Attach_ figures as *.pdf
files, for instance.
That being said, to answer your question, my advice (and that of many
others) is: DON'T. Error bars on boxplots (such plots are sometimes
known as "dynamite plots" or "detonator plots") are considered by the
cognoscenti to be abominations.
cheers,
Rolf
--
Technical Editor ANZJS
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276