CODEWITHSUNDEEP
rstatisticsplot
neversaint
neversaint

Reputation: 63984

How to Superimpose Multiple Density Curves Into One Plot in R

I have a data that looks like this.

And I intend to create multiple density curve into one plot, where each curve correspond to the unique ID.

I tried to use "sm" package, with this code, but without success.

library(sm)
dat <- read.table("mydat.txt");
plotfn <- ("~/Desktop/flowgram_superimposed.pdf");
pdf(plotfn);

sm.density.compare(dat$V1,dat$V2, xlab = "Flow Signal")
colfill <- c(2:10);
legend(locator(1), levels(dat$V2), fill=colfill)

dev.off();

Please advice what's the right way to do it or if there is alternative way to do it?

I am trying to get this kind of plot at the end. figure http://img524.imageshack.us/img524/2736/testl.png

Upvotes: 7

Views: 20599

Answers (4)

Steve Lianoglou
Steve Lianoglou

Reputation: 7263

I found myself needing to do this a lot when looking at microarray data, so I rolled this up as part of a library of utility code that I keep on github: ARE.utils, specifically the plot.densities function.

It uses base graphics so you can take inspiration from that function to create your own, or just take it whole-sale (but it relies on some other functions in that library):

  1. create.densities, which converts a list/matrix/etc of data to a list of densities; and
  2. match.dim function (which converts dimension "names" into numeric axes).

(You can, optionally, install the entire package, but I make no promises that I the functions in there won't change in some backwards incompatible way).

It's not hard to write your own such function, but just make sure you have the function pick the correct range on the axes and stuff. Anyway, you would then use the code like this:

library(ARE.utils)
# Create a matrix dataset with separate observations in columns
dat <- matrix(c(rnorm(100), rnorm(100, mean=3), 
                rnorm(100, mean=3, sd=2)),
              ncol=3)
# Plot them
plot.densities(dat, along='cols')

That would create three different density plots on the same axis with their own colors.

Upvotes: 1

Abhijit
Abhijit

Reputation: 625

You can also solve this using the lattice package.

require(lattice)
dnow <- read.table('http://dpaste.com/88561/plain/')
densityplot(~V1, groups=V2, data=dnow)

Upvotes: 4

Paolo
Paolo

Reputation: 2815

Using base graphics in a spaghetti code fashion:

plot.multi.dens <- function(s)
{
junk.x = NULL
junk.y = NULL
for(i in 1:length(s))
{
junk.x = c(junk.x, density(s[[i]])$x)
junk.y = c(junk.y, density(s[[i]])$y)
}
xr <- range(junk.x)
yr <- range(junk.y)
plot(density(s[[1]]), xlim = xr, ylim = yr, main = "")
for(i in 1:length(s))
{
lines(density(s[[i]]), xlim = xr, ylim = yr, col = i)
}
}
dnow <- read.table("http://dpaste.com/88561/plain/")
library(sqldf)
x <- unlist(sqldf("select V1 from dnow where V2==0"))
y <- unlist(sqldf("select V1 from dnow where V2==1"))
z <- unlist(sqldf("select V1 from dnow where V2==2"))
plot.multi.dens(list(x,y,z))
library(Hmisc)
le <- largest.empty(x,y,.1,.1)
legend(le,legend=c("x","y","z"), col=(1:3), lwd=2, lty = 1)

Upvotes: 3

Eduardo Leoni
Eduardo Leoni

Reputation: 9050

Try using ggplot2:

dnow <- read.table("http://dpaste.com/88561/plain/")
library(ggplot2)
qplot(V1, colour=factor(V2), data=dnow, geom="density")

Upvotes: 11

Related Questions