How may I reproduce the `Distributions.jl` logo?

Hey,

I’m trying to reproduce the Distributions.jl logo. Is there somewhere a code I can use to reproduce this logo ? I cannot retrace back the logo further than this PR : https://github.com/JuliaStats/Distributions.jl/pull/640, however it does not say where it came from (actually it was moved).
@ararslan you were there at the time, do you remember ?

I have the following :

using Plots
using Plots.PlotMeasures
using StatsBase



function make_plot(data)
    return scatter(data[1,:],
    data[2,:],
    color=colors,
    msw=0,
    # ms=2,
    α=0.2,
    label=:none,
    size=[600,600],
    bottom_margin=0mm, 
    left_margin = 0mm, 
    right_margin = 0mm, 
    top_margin = 0mm,
    axis=nothing,
    ticks=nothing,
    border=:none);
end
random_noise(N) = randn((2,N)) ./ 3.2
function perfect_noise(N)
    rez = zeros((2,N))
    for i in 1:N
        x = exp(rand()*2π*im)*rand()/1.3
        rez[:,i] .= x.re,x.im
    end
    return rez
end

mynoise = random_noise

N = 10000
c1 = [cos(3π/6), sin(3π/6)] .+ mynoise(N)
c2 = [cos(7π/6), sin(7π/6)] .+ mynoise(N)
c3 = [cos(11π/6),sin(11π/6)] .+ mynoise(N)
colors = [repeat([:green],N)...,
          repeat([:red],N)...,
          repeat([:purple],N)...]
data = hcat(c1,c2,c3)
p1 = make_plot(data)

But this is not quite perfect, compared to the true logo : https://github.com/JuliaStats/Distributions.jl/blob/master/docs/src/assets/logo.png

Hence, the noise seems not to be Gaussian. What else ?

1 Like

There’s some discussion here - https://github.com/JuliaStats/Roadmap.jl/issues/22

1 Like

Here’s a quick attempt at reproducing it:

using Luxor

@draw begin
	background("black")
	translate(0, 30)
	setopacity(0.7)
	N = 3000
	for (n, pt) in enumerate(ngon(O, 120, 3, π/6, vertices=true))
		sethue([Luxor.julia_red, Luxor.julia_green, Luxor.julia_purple][n])
		box.(pt .+ 35 .* Point.(randn(N), randn(N)), 6, 6, :fill)
		strokepath()
	end
end

stats

but I couldn’t get all the dots to be in the same place as the original … :rofl::rofl::rofl:

Yep me neither !

Thanks, thats pretty.

1 Like