plot - Plotting Probability Density Heatmap Over Time in R -

let's have output of monte-carlo simulation of 1 variable on several different iterations (think millions). each iteration, have values of variable @ each point in time (ranging t=1 t=365).

i produce following plot: each point in time, t, on x axis , each possible value "y" in given range, set color of x,y "k" "k" count of how many observations within vicinity of distance "d" x,y.

i know can make density heatmaps 1d data, there package doing on 2 dimensions? have use kriging?

edit: data structure matrix.

                                     data matrix                                        day number              [,1]    [,2]         [,3]      [,4]       [,5]      ... [,365] iteration    [1,]    0.000213   0.001218    0.000151   0.000108  ... 0.000101              [2,]    0.000314   0.000281    0.000117   0.000103  ... 0.000305              [3,]    0.000314   0.000281    0.000117   0.000103  ... 0.000305              [4,]    0.000171   0.000155    0.000141   0.000219  ... 0.000201               .               .               .      [100000000,]    0.000141   0.000148    0.000144   0.000226  ... 0.000188 

i want to, each "day" have pixels running vertically across "day" represent probability density of iteration's values day in color. result should heatmap.

here 1 solution think after.

1. generate data.

   mydata <- mapply(rnorm, 1000, 200, mean=seq(-50,50,0.5))

this matrix 1000 rows (observations) , 201 time points. in each time point mean of data there shifts gradually -50 50. 0.5 each time.

2. get densities.

   mydensities <- apply(mydata, 2, density, from=-500, to=500)

this give list of densities each column. in order them plottable side side specified ranges (from -500 500) manually.

3. obtain density values list.

   ys <- sapply(mydensities, "[", "y")

this again list. need matrix that.

4. get matrix list.

   img <- do.call(cbind, ys)

this combines ys elements column.

5. plot.

   filled.contour(x=1:ncol(img), y=mydensities[[1]]$x, t(img))

i use filled.contour that. can around other 2-d plot functions. used values obtained densities d[[1]]$x.

and here result:

the shift -50 50 visible.

not sure if can work millions of time points. plotting million makes little sense since in case limited number of pixels. kind of pre-processing might necessary.