CODEWITHSUNDEEP
r
Bfu38
Bfu38

Reputation: 1141

plot a "list" of densities

I have a data.frame composed of 10 columns and 50 rows. I computed density function, column by column using the apply function. Now I would like to plot the densities I computed at once.

In other words, instead of plotting...

plot(den[[1]]) 
plot(den[[2]]
plot(den[[3]]

...I would like to plot all the densities at once.

I suppose I have to use the lapply function and probably I have to write a specific function to do this. Can anyone help me?

Upvotes: 6

Views: 1392

Answers (3)

agstudy
agstudy

Reputation: 121568

latticeExtra has very handy function, marginalplot 

marginal.plot(DF,layout=c(2,2))

enter image description here

Upvotes: 1

Marek
Marek

Reputation: 50704

I think that lattice package could be handy:

Following example from Jilber:

set.seed(001)
DF <- data.frame(matrix(rnorm(400, 100, 12), ncol=4)) # some data
DF <- stack(DF) # to long form

library(lattice)
densityplot(~values|ind, DF, as.table=TRUE)
# or
densityplot(~values, groups=ind, DF)

Results are:

Separate densities

and

Combined densities

Upvotes: 1

Jilber Urbina
Jilber Urbina

Reputation: 61154

Maybe this would be helpful

set.seed(001)
DF <- data.frame(matrix(rnorm(400, 100, 12), ncol=4)) # some data
den<-apply(DF, 2, density) # estimating density


par(mfrow=c(2,2))
sapply(den, plot) # plot each density
par(mfrow=c(1,1))

Which gives...

enter image description here

Giving some names:

par(mfrow=c(2,2))
for(i in 1:length(den)){
  plot(den[[i]], 
       main=paste('density ', i))
}
par(mfrow=c(1,1))

enter image description here

If you just don't want all plots in one output you may want to leave par(mfrow=c(2,2)) out and just run sapply(den, plot)

Edit: Answer for your second question (in the comment)

plot(den[[1]], ylim=c(0,.04), 
     main='Densities altogether') # plot the first density
for(i in 2:length(den)){          # Add the lines to the existing plot
  lines(den[[i]], col=i)          
}

enter image description here

Here is useful using legend function to add a legend

legend('topright', paste('density ', 1:4), col=1:4, lty=1, cex=.65)

Upvotes: 7

Related Questions