Hi:
Here's a fairly basic ggplot2 version with a few warts in it:
# Generate means
means <- rpois(30, 10)
# Create 30 Gaussian random samples of size 100, using means to
# define the population mean for each sample (via sapply)
df <- data.frame(ds = rep(1:30, each = 100),
x = as.vector(sapply(means, function(x) rnorm(100, means[x]))) )
library(ggplot2)
# Two versions of density plot:
# Version 1: no fill color, just curves
h <- ggplot(df2, aes(x = x, colour = ds))
h + geom_density() + facet_grid(ds ~ ., as.table = FALSE) +
opts(axis.text.y = theme_blank())
# as.table = FALSE allows you to read upward from data set 1 to 30
# Version 2: add fill color
g <- ggplot(df2, aes(x = x, colour = ds, fill = ds))
g + geom_density() + facet_grid(ds ~ ., as.table = FALSE) +
opts(axis.text.y = theme_blank())
Some problems that arose:
(1) I couldn't figure out how to set a range of colors beyond the
default. I tried several things with scale_colour_*, but one
problem or another arose. Ideally, I'd like to be able to set
the color by mean value, but that would require some thinking
(OMG!) Some features of scales are still a bit beyond me...
(2) I could get rid of the y-axis tick labels, but couldn't get
rid of both the labels and the ticks. I was also hoping to use
theme_bw() using just the horizontal grid lines, but no. I just
thought of a couple ideas to play with on that front., though,
so I'll continue to toy with it for a bit.
All in all, for a few lines of code, it's a pretty decent plot, even
though it needs some work. Thank Hadley for that...
HTH,
Dennis
On Sun, Sep 12, 2010 at 1:17 PM, Karl Brand <k.br...@erasmusmc.nl> wrote:
> Cheers!
>
> All excellent, runable examples helping me progress quickly.
>
> Being more a qualitative plot, the y-axis is less important. But it did get
> me thinking-
>
> Coloring each of the plotted lines, say 'altitude colors' like the classic
> volcano example to reflect the (scaled) values the lines represent might be
> effective at representing individual y-axis magnitude for each line. Perhaps
> gray background at least.
>
> Maybe if there were examples of the lines() func. using colors dependent on
> the y-value? Or some one already made a function for achieving this? (Google
> didnt return anything obvious for me...yet).
>
> Sincere thanks for help thus far, and welcome any further pointers,
>
> Karl
>
>
>
> On 9/12/2010 4:50 PM, Duncan Murdoch wrote:
>
>> On 12/09/2010 10:12 AM, Karl Brand wrote:
>>
>>> Esteemed useRs and developeRs,
>>>
>>> I need to create a '3D line plot' (proper name?) of which an excellent
>>> example can be viewed here:
>>>
>>> http://cococubed.asu.edu/images/87a/images/unknown_pleasures.jpg
>>>
>>> I have some experience using the rgl package to create 3D PCA plots,
>>> but have no idea where to start for an image like this.
>>>
>>> I'd really appreciate suggestions & help on how might achieve this
>>> using R,
>>>
>>
>> I wouldn't use rgl for that: it's really a flat 2D plot, without
>> perspective, shading or anything else that 3D plots have other than
>> foreground objects hiding background ones. You can draw it by using
>> polygon() to hide the background then lines() to draw the lines.
>>
>> Here's an example using regular 2d graphics:
>>
>>
>> n <- 60
>> m <- 50
>> x <- seq(-4,4, len=m)
>>
>> # Make up some fake y data
>>
>> y <- matrix(NA, n, m)
>> for (i in 1:n) y[i,] <- dnorm(x)*runif(m, 0.5,1)
>>
>> par(bg="black")
>> yrange <- range(c(y, y+n/20))
>>
>> plot(x, x, type="n", axes=FALSE, bg="black", ylim=yrange)
>>
>> for (i in n:1) {
>> y1 <- c(y[i,] + i/20, 0, 0)
>> x1 <- c(x, x[m], x[1])
>> polygon(x1,y1,col="black")
>>
>> lines(x, y[i,] + i/20, col="white")
>>
>> }
>>
>> The problem with this kind of plot is that it's almost impossible to
>> display a usable vertical scale.
>>
>> Duncan Murdoch
>>
>
> --
> Karl Brand
> Department of Genetics
> Erasmus MC
> Dr Molewaterplein 50
> 3015 GE Rotterdam
> T +31 (0)10 704 3457 |F +31 (0)10 704 4743 |M +31 (0)642 777 268
>
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
