Enh scale collections with size

[tacaswell]

This was prompted by http://stackoverflow.com/questions/32444037/how-can-i-create-many-thousands-of-circles#32446502

I suspect that this is not the right way to achieve this (as it is mostly broken under log transform and the notion of 'area' under blended transform is confused)

but, with this patch

import matplotlib.pyplot as plt
import matplotlib.collections as mc
import numpy as np

ax = plt.gca()
cc = mc.CircleCollection(sizes=.001 * np.pi*np.random.rand(10000)**2,
                         offsets=np.random.rand(10000, 2),
                         transOffset=ax.transData,
                         facecolors='rgb')
cc.set_size_transform(ax.transData)
ax.add_collection(cc)


ax.set_aspect('equal')

Does the 'right thing' and renders relatively snappily (0.1-0.2 s).

[tacaswell]

This still needs docs + tests even if this is the right direction to go.

ping @mdboom

[jklymak]

Closing for lack of progress... Happy to re-open if this approach should be followed up...