Custom plotting utilities for matplotlib

Some useful little tidbits for plotting in matplotlib. Contains tools for plotting timeseries (smoothed and annotated autocorrelation plots) and for thinning down a collection of a large number of points in $n$ dimensions based off of a minimum distance between any two points (rather like thinning seedlings in a planting bed).

This latter tool works best for plotting scatterplots in 2D, potentially with millions of datapoints, with either linear or log axes, where a set of representative points (and weights denoting how many actual points they represent) is desired, rather than a density-histogram visualization (e.g. datashader). The representative-sample visualization may be more useful for showing the overall spread of the data while preserving a sense for how tightly they cluster about the central distribution.

Also contains the four-colour "Dark2" palette from colorbrewer, which should be readable for most people with some type of colour vision deficiency. Note that this palette is now available directly in matplotlib:

from matplotlib import pyplot as plt
print(plt.cm.Dark2.colors)

Copied from my other repo.