Problem with plotting two matrices and using a loop

[silke]

Hi everyone,
I am trying to "translate" a plot into a ggplot to keep my graphics
consistent, but I am facing some problems and hope somebody can help.
I would like to plot two matrices against each other, like in a
scatterplot. This works easily with plot, for example:
x = matrix(rnorm(100),nrow=10)
y = matrix(rnorm(100),nrow=10)
plot(x,y)
or as well in a loop:
plot(x[1,], y[1,],type="n") # maybe with xlim,ylim argument
for (i in 1:10) {
points(x[i,], y[i,]) }

However, I don't get this to work with ggplot.
Obviously, I cannot just plot the matrices as above like "qplot(x,y)"
as I get an error message, I do not really understand why.
I have tried
xy = data.frame(x,y)
m = ggplot(xy)
m + for (i in 1:10) {
geom_point(data=dafr,aes(x=x[,i],y=y[,i]))
}
But then only the last layer (i=10) of the loop is plotted. If I would
explicitly write all ten loops it works as I want, but my matrices are
actually 160x160, so I would really like to find a better solution.
I have tried something with print() because I've read that it is
needed within loops, but it did not work either.
Any hint would be very much appreciated!

Thanks, Silke.

[silke]

Hallo again,
I just saw a little mistake in my yesterday's example of what I tried
that didn't work, i meant:

xy = data.frame(x,y)
m = ggplot(xy)
m + for (i in 1:10) {

geom_point(data=xy,aes(x=x[,i],y=y[,i]))
}
But still it does not work and only the last layer is printed -
doesn't anybody have a solution for me in this?
Thank you very much in advance,
Silke Schmidt.

[hadley wickham]

On Thu, Jan 22, 2009 at 4:13 AM, silke <sil...@gmail.com> wrote:
>
> Hallo again,
> I just saw a little mistake in my yesterday's example of what I tried
> that didn't work, i meant:
> xy = data.frame(x,y)
> m = ggplot(xy)
> m + for (i in 1:10) {
> geom_point(data=xy,aes(x=x[,i],y=y[,i]))
> }
> But still it does not work and only the last layer is printed -
> doesn't anybody have a solution for me in this?
> Thank you very much in advance,

Have a look at what this code does:

a <- 1
a + for(i in 1:10) {a}

A for loop returns the last value. You want something like:

for (i in 1:10) {

print(m + geom_point(data=xy,aes(x=x[,i],y=y[,i])))
}

Regards,

Hadley

--
http://had.co.nz/

[silke]

Thank you very much for your response, although the loop did not
really work for me as it returns 10 different plots, one of each
layer, and what I wanted is to have all layers in the same plot.
However I found another solution with the help of a friend:

x = as.numeric(matrix(rnorm(100),nrow=10))
y = as.numeric(matrix(rnorm(100),nrow=10) )

xy = data.frame(x,y)
m = ggplot(xy)

m + geom_point(aes(x=x,y=y))

Seems to be a problem of the data type...

Kind regards,
Silke.

On 22 Jan., 14:13, hadley wickham <h.wick...@gmail.com> wrote:

[hadley wickham]

Ok, in that case you want the following in the body of your loop:

m <- m + geom_point(aes(x=x[,i],y=y[,i]))

Hadley

--
http://had.co.nz/

[silke]

Hi,

I am sorry to bother again with this issue, but actually I avoided the
above loop problem by converting the data using as.numeric().
I don't know if I missed something in your answer above, but I didn't
get your solution to work, the result were (as above) 10 different
plots of each layer and not all 10 layers in one plot.
Now I need a similar plot with lines insted of points, and my
as.numeric solution does not work because I do not want all points to
be connected by a line, but:
The data points are in a 50x160 matrix, and I need each column of the
matrix to be one line (of length 50) in the plot which thus contains
160 single lines, plotted e.g. just against the numbers 1:50.
I also put the data into a list of length 160, with each list object
containing the vector I want to be one line, and tried to use melt (),
but I didn't get it to work neither.

I guess the solution must be quite simple but I'm stuck...

A simple example with a plot solution would be:
x = matrix(rnorm(50*160),ncol=160)
for (i in 1:160) x[,i]=x[,i]+i
plot(seq(1,50),x[,1],ylim=range(x),type="l")
for (i in 1:160) lines(seq(1,50),x[,i])

Can anybody help me finding a ggplot solution?

Many thanks in advance,
Silke.

On 22 Jan., 18:00, hadley wickham <h.wick...@gmail.com> wrote:
> Ok, in that case you want the following in the body of your loop:
>
> m <- m + geom_point(aes(x=x[,i],y=y[,i]))
>
> Hadley
>
>
>
> On Thu, Jan 22, 2009 at 10:04 AM, silke <silk...@gmail.com> wrote:
>
> > Thank you very much for your response, although the loop did not
> > really work for me as it returns 10 different plots, one of each

> > layer, and what I wanted is to have alllayersin the same plot.

[baptiste auguie]

Hi,

If I understand correctly, you can try this:

y <- matrix(rnorm(50*20),ncol=20)

mdf <- data.frame(y)
mdf$x <- seq(1, nrow(y))

qplot(x, value, data=melt(mdf, id="x"), colour= variable , geom="line")

(I replaced 160 by 20 as it was quite slow on my machine)

note that in base graphics you could use matplot(1:50, y, lty=1)


Hope this helps,

baptiste

---

From: silke <sil...@gmail.com>
To: ggplot2 <ggp...@googlegroups.com>
Sent: Thursday, 5 February, 2009 15:19:00
Subject: Re: Problem with plotting two matrices and using a loop

[silke]

Yes, that helps, thank you very much!

On 5 Feb., 17:32, baptiste auguie <bapt_4...@yahoo.fr> wrote:
> Hi,
>
> If I understand correctly, you can try this:
>
> y <- matrix(rnorm(50*20),ncol=20)
>
> mdf <- data.frame(y)
> mdf$x <- seq(1, nrow(y))
>
> qplot(x, value, data=melt(mdf, id="x"), colour= variable , geom="line")
>
> (I replaced 160 by 20 as it was quite slow on my machine)
>
> note that in base graphics you could use matplot(1:50, y, lty=1)
>
> Hope this helps,
>
> baptiste
>
> ________________________________

> From: silke <silk...@gmail.com>