Is it possible to assign a colourmap for edgecolors in matplotlib?

Question

Assuming one needs to change the edge colour of a matplotlib marker depending on some variable, is it possible to assign a some sort of discrete colour-map for the edge colour of the marker? This is similar to changing the face-colour of the marker by cmap.

When showing limits using arrows outside a plot's range, I cannot seem to vary the arrow colour depending on another variable. eg: in the code below the colour of the arrow doesn't change as a function of z.

plt.scatter(x,y, c=z, marker=u'$\u2191$', s=40,cmap=discrete_cmap(4, 'cubehelix') )

Answer 1

You can do this using the edgecolors argument to scatter.

You need to make a list of colours to feed to scatter. We can do this using your chosen colormap and a Normalize instance, to rescale to z function to the 0-1 range.

I'm assuming your discrete_cmap function is something like the one linked here.

import matplotlib.pyplot as plt
import matplotlib.colors as colors
import numpy as np

# def discrete_cmap() is omitted here...

# some sample data
x = np.linspace(0,10,11)
y = np.linspace(0,10,11)
z = x+y

# setup a Normalization instance
norm = colors.Normalize(z.min(),z.max())

# define the colormap
cmap = discrete_cmap(4, 'cubehelix')

# Use the norm and cmap to define the edge colours
edgecols = cmap(norm(z))

# Use that with the `edgecolors` argument. Set c='None' to turn off the facecolor
plt.scatter(x,y, edgecolors=edgecols, c = 'None', marker='o', s=40 )

plt.show()