Set linetype manually for each line
thomas
is there a way to manually set which linetype ggplot2 should use for a
specific line?
I have 4 lines and set the linetype simply by saying:
ggplot(data,aes(x=ID,y=value,group=Section)) +
geom_line(aes(linetype=Section,colour = Section))
The problem is that ggplot kind of selects the wrong linetypes for the
lines. Attached is the end result.
http://shareimage.org/images/w58rjvo5052jb5p3pie4.jpg
I would rather have Total having the linetype of Abstract for
example.
Anybody knows how that could be done?
Thank you,
Thomas
Joshua Wiley
I see two options. You can either manually specify the linetype for
each level of your factor, OR you can reorder your factor to match
your desired line types (obviously somewhat less flexible, but can be
easier code wise). Example code follows.
Cheers,
Josh
library(ggplot2)
## data
dat <- data.frame(x=1:10, y=1:10,
group = factor(rep(letters[1:5], each = 2)))
## default lines
ggplot(data = dat, aes(x = x, y = y, linetype = group)) +
geom_line()
## manually set liens
ggplot(data = dat, aes(x = x, y = y, linetype = group)) +
geom_line() +
scale_linetype_manual(name = "Linetype",
value = c("a" = 2, "b" = 3, "c" = 4, "d" = 1, "e" = 5))
## reorder factor so non-alphabetical
dat$group <- factor(dat$group, levels = c("d", "a", "b", "c", "e"))
## new default lines
ggplot(data = dat, aes(x = x, y = y, linetype = group)) + geom_line()
> --
> You received this message because you are subscribed to the ggplot2 mailing list.
> Please provide a reproducible example: http://gist.github.com/270442
>
> To post: email ggp...@googlegroups.com
> To unsubscribe: email ggplot2+u...@googlegroups.com
> More options: http://groups.google.com/group/ggplot2
>
--
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/
thomas
thank you very much.
You are using factor and are reordering the levels in your last
method.
If i don't provide a factor, would ggplot2 internally create a factor
and use its levels to create a mapping between the linetype and the
unique values? And by handing ggplot the factor directly you make it
skip this step?
Thomas
Joshua Wiley
I just sort of assumed you were using a factor, I guess. AFAIK things
like the scale training and mapping is done using factors for discrete
data. When factor() is called on data, the levels are, by default,
alphabetical, but if you have already created factor class data and
specified a different ordering for levels than those will be kept.
The linetype and legend order are, again AFAIK, determined by the
levels of the factor.
Take that with some caution though, I have not actually watched all
the different steps between the call and the final graph.
HTH,
Josh
James McCreight
ecm <- melt(ec_scaled, id = "date")
qplot(date, value, data=ecm, geom="line", linetype=facord)
ecm$char<-as.character(ecm$facord)
qplot(date, value, data=ecm, geom="line", linetype=char)
******************************************************************************
James McCreight mccreigh at colorado.edu
NASA Postdoctoral Fellow
cell: (831) 261-5149
thomas
first time I read about factor() I didn't really understand what it is
good for.
Thanks for the explanation, that clears things up quite a bit.
Thanks,
Thomas
thomas
thank you, I will play around with factor() from now on.
Best,
Thomas
On Nov 16, 6:11 pm, James McCreight <mccre...@colorado.edu> wrote:
> it appears that the scale_linetype() is applying factor() to its grouping
> variable, which defaults to alphabetical order. but i would play it save and
> make it a factor on your own. generally, when "discretizing" variables, the
> burden falls on the user. but as you see, this affords the user some amount
> of flexibility. (same goes with using cut() to get factors for discrete
> scales.)
>
>
> ec_scaled <- data.frame( date = economics$date, rescaler(economics[,
> -(1:2)], "range") )
> ecm <- melt(ec_scaled, id = "date")
>
> qplot(date, value, data=ecm, geom="line", group=variable)
>
>
> ecm$char<-as.character(ecm$facord)
> qplot(date, value, data=ecm, geom="line", linetype=char)
>
>
>
>
>
>
>
>
>
>
> > Hi Thomas,
>
> > I just sort of assumed you were using a factor, I guess. AFAIK things
> > like the scale training and mapping is done using factors for discrete
> > data. When factor() is called on data, the levels are, by default,
> > alphabetical, but if you have already created factor class data and
> > specified a different ordering for levels than those will be kept.
> > The linetype and legend order are, again AFAIK, determined by the
> > levels of the factor.
>
> > Take that with some caution though, I have not actually watched all
> > the different steps between the call and the final graph.
>
> > HTH,
>
> > Josh
>
> > > Hi Joshua,
>
> > > thank you very much.
> > > You are using factor and are reordering the levels in your last
> > > method.
> > > If i don't provide a factor, would ggplot2 internally create a factor
> > > and use its levels to create a mapping between the linetype and the
> > > unique values? And by handing ggplot the factor directly you make it
> > > skip this step?
>
> > > Thomas
>
> > > --
> > > You received this message because you are subscribed to the ggplot2
> > mailing list.
> > > Please provide a reproducible example:http://gist.github.com/270442
>
> > > To post: email ggp...@googlegroups.com
>
> > --
> > Joshua Wiley
> > Ph.D. Student, Health Psychology
> > University of California, Los Angeles
> >http://www.joshuawiley.com/
>
> > --
> > You received this message because you are subscribed to the ggplot2 mailing
> > list.
> > Please provide a reproducible example:http://gist.github.com/270442
>
> > To post: email ggp...@googlegroups.com
>
> --
> -
> *************************************************************************** ***
> James McCreight mccreigh at
> edu