[R] graphics
Deepayan Sarkar
deepayan.sarkar at gmail.com
Tue Aug 30 20:49:49 CEST 2005
On 8/30/05, Karsten Rincke <rincke at physik.uni-kassel.de> wrote:
> Hello,
> I guess a have a very simple problem though up to now couldn't solve it:
> I want to plot two datasets wihtin one plot like plot(x) provides it for
> one dataset(type="b" that is: points connected by lines).
> Example data 'x':
> Befragung1 Befragung2 Befragung3 Geschlecht
> 2.25 2.34 1.78 weiblich
> 1.34 3.45 2.23 maennlich
> The two rows of the example above form two datasets. Now I'm looking for
> something like plot(~x |Geschlecht, ... ): X-axis containing
> Befragung1...3, y-axis containing the values given by the matrix 'x',
> points within one dataset (row of 'x') connected by lines.
> I already tried various possibilities provided by 'lattice' or 'grid' -
> but there seems to be a basic misunderstanding on my side so that I
> could not produce any good result.
> Would by very happy to get some hints!
R likes data vectors to be columns, not rows. With your data, you
could try
matplot(t(mydata[, 1:3]), type = 'b')
but it would probably make more sense to have the data in a different
format to begin with. It could either be in the `wide' format:
Befragung weiblich maennlich
1 2.25 1.34
2 2.34 3.45
3 1.78 2.23
or the `long' format:
Befragung y Geschlecht
1 2.25 weiblich
2 2.34 weiblich
3 1.78 weiblich
1 1.34 maennlich
2 3.45 maennlich
3 2.23 maennlich
With the wide format you could use matplot as before, or use xyplot
from lattice:
mydata <-
read.table(textConnection("
Befragung weiblich maennlich
1 2.25 1.34
2 2.34 3.45
3 1.78 2.23"), header = TRUE)
xyplot(maennlich + weiblich ~ Befragung,
data = mydata,
type = 'b',
auto.key = TRUE)
With the long format, you could similarly do:
mydata <-
read.table(textConnection("
Befragung y Geschlecht
1 2.25 weiblich
2 2.34 weiblich
3 1.78 weiblich
1 1.34 maennlich
2 3.45 maennlich
3 2.23 maennlich
"), header = TRUE)
xyplot(y ~ Befragung, data = mydata, groups = Geschlecht,
type = 'b', auto.key = TRUE)
Hope that helps,
-Deepayan
More information about the R-help
mailing list