Plots.jl: Sparsify Data

[cortner]

Sometimes I have millions of data points to plot, e.g., with scatter. This is extremely inefficient so I first sparsity them by throwing away points that are close within some tolerance of already existing points.

Surely this must come up on a regular basis. Maybe this is something that should be left with the user, but I thought I'd mention it as a possible feature for Plots?

[mkborregaard]

Anyone could quite easily define a sparseplot userplot recipe that does this, and PR it on PlotRecipes (rather than Plots I'd say). I'm sure it would be useful for other people in your situation.

[cortner]

that makes perfect sense - I haven't actually properly looked at PlotRecipes, but will do when I have the time.

[piever]

This recipe has value, but I wanted to mention that there is another approach in OnlineStats, where you can give a partition of the data and it will plot for example the scatter of some summary statistics on that partition. The link above has a few examples, see if they satisfy your usecase.

[cortner]

Interesting - I haven’t thought about visualising my data that way, in my specific case it is not obvious that it would represent what the data means, but thank you for pointing t out.