How to plot a sequence of probability density functions as lines in (pseudo) 3D?

Specific Domains Visualization
, , , ,
1

Given a set of PDFs f1(x), f2(x) etc., I’d like to plot something like this:

1556×1167 286 KB

The closest method I found is density from Makie, but it doesn’t work with functions or their values per se.

How can I make such a plot?

2

Based on the code here, perhaps something like this:

using GLMakie
using Distributions

fig = Figure()
ax = Axis3(fig[1,1])
yr = 1975:2015
xs = range(-15,10,length=200)
μs = range(-14,9,length=length(yr))
σs = range(1.0, 5.0, length=length(yr))

for i ∈ eachindex(yr)
    zs = pdf.(Normal(μs[i], σs[i]), xs)
    band!(ax, Point3f.(tuple.(xs, yr[i], 0.0)), Point3f.(tuple.(xs, yr[i], zs )); color = zs, alpha = 0.75)
end
xlims!(ax, -15, 10)
ylims!(ax, 1975, 2015)
zlims!(ax, 0.0, 0.4)
fig

pdfs
pdfs800×600 69.4 KB

3

Linking related thread for more ideas.

1 Like
4

Thanks! Is there an short way to render it with lines? Changing band to lines results in an error.

5

It has some nice examples, thanks! I’ve learnt it’s called the waterfall plot.

1 Like
6

This looks like a plot I’ve seen before. Growth at Risk paper? Are you in econ?

7

Sure, replace the line that has band! with the line

lines!(ax, xs, fill(yr[i],length(xs)), zs; color = :black, lw = 1.0)

You can use both too, if you like if you put the the above command below the band! line.

1 Like
8

It the first example I found. I simply need to visualize and compare some PDFs.

1 Like