Closed ararslan closed 7 years ago
Good idea, but the curve should be more like a normal distribution, and we should have a SVG version.
Something like this might be cool too:
Isnt that a bit too much Freudian innuendo¿ ^^
😖 Maybe to ameliorate that we could move the means of the red and purple distributions away from the center and decrease their variances a bit?
@mschauer I learned a lot about you today!
@ararslan Feel free to play with that bit of code and come up with something else!
@tbreloff I actually can't seen to get Plots to work under 0.3, 0.4, or 0.5. But I'm fine with whatever, it was just a suggestion. 😄
@ararslan Do you mind opening a Plots issue or chatting in gitter with more details?
Modification of Tom's image.
I actually like @ararslan's original idea better.
The colors are a bit too flashy...
Cool!
I like this one. Maybe without the axes?
On Fri, May 13, 2016 at 9:12 AM, Kristoffer Carlsson < notifications@github.com> wrote:
Cool!
— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub https://github.com/JuliaStats/Roadmap.jl/issues/22#issuecomment-219038801
Yes there should be noise.
I like this one as well. But is it too similar to Tom's image for JuliaML?
Bump for this discussion. @cstjean's suggestion looks great to me but I agree with @andreasnoack that the axes should be removed.
I made the axes transparent and cropped it manually. I couldn't figure out how to make the background transparent, but that could be done manually too.
If anyone else wants to tweak it further:
using Plots
plotly()
N = 2000
w = 0.3
al = 0.7
m = (:circle, 2.5, 0.4, stroke(0))
l = 2.5
scatter(randn(N)*w, randn(N)*w+2, color=:green, size=(200,200), legend=false, alpha=al, marker=m, grid=false, foreground_color=RGBA(0.0,0.0,0.0,0.0), xlims=(-l,l), ylims=(1-l, 1+l))
scatter!(randn(N)*w-1, randn(N)*w, color=:red, alpha=al, legend=false, marker=m)
scatter!(randn(N)*w+1, randn(N)*w, color=:purple, alpha=al, legend=false, marker=m)
I don't know how much the transparent background actually matters; the JuliaMath logo's background is white, for example.
Modified @cstjean's to reduce the height so that the dot centers form an equilateral triangle and made the background transparent.
+1
I think it would be nice if JuliaStats could have a distinguishing logo, just as JuliaDB and others do, rather than the same logo as JuliaLang. Here's my proposal, poorly made in the course of about 5 minutes in an MS Paint clone:
It's just the Julia dots with the normal distribution overlaid. Not super exciting but I figured that the normal distribution is pretty iconic and could work well in this scenario.
Thoughts? I'm certainly no graphic designer; this is just an idea. (Though if people do like this, I'd request that someone with decent image software make one that looks less sloppy, especially if we need transparency.)