quantified-uncertainty / squiggle

An estimation language
https://squiggle-language.com
MIT License
148 stars 22 forks source link

Scatter plot default point color makes it undiscernable from dirt on the screen #3269

Open cab404 opened 1 month ago

cab404 commented 1 month ago

Description:

image

Steps to reproduce:

https://www.squiggle-language.com/playground?v=0.9.4#code=eNqrVkpJTUsszSlxzk9JVbJSSlGwVQjIyS%2FRK05OLClJLdKojslTUKhwySwusVIITswtyEkNTi3RSyvKz%2FUBimmACL3ijNK0tJxUCKe0ICRfw0DH0MBAU1NTB6S7Er9uZA0xebWaSjpKxRn55cGlubmJRZVKViVFpam1APKPORo%3D

d = Plot.scatter({
  xDist: SampleSet.fromList(List.shuffle(List.upTo(0,100))),
  yDist: SampleSet.fromList(List.upTo(0,100))
})

Expected behavior:

Other color for points

What I got instead:

Off-white point color

OAGr commented 1 month ago

Yep, fair. One thing here is that the system is optimized for 1k+ points, when we want them to be fainter. We can adjust to have opacity be proportional to number

cab404 commented 1 month ago

Yep, fair. One thing here is that the system is optimized for 1k+ points, when we want them to be fainter. We can adjust to have opacity be proportional to number

IMO it depend on a dataset in more ways than just number of points — e.g you can have really uniform coverage of a whole plot with points, and you are trying to spot places with no points.

Plot.style(plot, {height?: int, colors?: int[], ...})?