Custom Matplotlib colourmap - like pgfplots

Question

Is it possible to get the Blue-Yellow-Red colourmap from pgfplots , as in this example https://tex.stackexchange.com/questions/39145/scaling-colormaps-with-pgfplots , in matplotlib ?

I have taken a look on how to define custom colourmaps in matplotlib, but it seems very difficult.

Answer 1

Defining a custom cmap is the way to go though. It's not that difficult but you must make sure you understand all the values in the color dict.

For each color you specify three columns of values. The first column is the position at which point a color occurs, 0 being the first and 1 the last. Since you want three colors, there are three evenly spaced positions defined 0, 0.5 and 1.0.

You want:

0.0: Blue

0.5: Yellow

1.0: Red

The second column defines the color 'up to' the position, the third column the color 'from/after' that position. If you want colors to gradually fade you should keep them the same. Specifying different colors allows the colormap to have sudden 'jumps' after a certain position.

To get blue at position 0.0, set red and green to 0 and blue to 1. To get yellow at position 0.5 set red and green to 1 and blue to zero. And so on...

import matplotlib as mpl
from matplotlib import pyplot as plt
import numpy as np

cdict = {'red': ((0.0, 0.0, 0.0),
                 (0.5, 1.0, 1.0),
                 (1.0, 1.0, 1.0)),
        'green':((0.0, 0.0, 0.0),
                 (0.5, 1.0, 1.0),
                 (1.0, 0.0, 0.0)),
         'blue':((0.0, 1.0, 1.0),
                 (0.5, 0.0, 0.0),
                 (1.0, 0.0, 0.0))}

custom_cmap = mpl.colors.LinearSegmentedColormap('mymap', cdict, 256)

plt.imshow(np.arange(36).reshape(6,6), cmap=custom_cmap, interpolation='none')

Answer 2

@RutgerKassies has a great explanation of how linear colormaps work in matplotlib in general. When you need colormaps with uneven transitions or "hard" cutoffs, you need to specify the full dict of threshold values.

However, there's a convenience "factory function" for simple colormaps with even transitions between colors: LinearSegmentedColormap.from_list (Actually, you can even make uneven transitions with it, just as long as they're not "hard" transitions.)

For example (to steal Rutger's example), we can do this with:

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

cmap = LinearSegmentedColormap.from_list('mymap', ['blue', 'yellow', 'red'])

plt.imshow(np.arange(36).reshape(6,6), cmap=cmap, interpolation='none')
plt.show()