Adjusting gridlines on a 3D Matplotlib figure

Question

I\'m getting ready for a presentation and I have some example figures of 3D matplotlib figures. However, the gridlines are too light to see on the projected images.

Answer 1

Unfortunately this doesn't seem to be exposed. Looking over the source, the key internal variable is call _AXINFO which we can override by careful subclassing.

Add this code after your figure is created, and style it with the dict custom_AXINFO:

from mpl_toolkits.mplot3d import Axes3D
import mpl_toolkits.mplot3d.axis3d as axis3d

# New axis settings
custom_AXINFO = {
    'x': {'i': 0, 'tickdir': 1, 'juggled': (1, 0, 2),
          'color': (0.00, 0.00, 0.25, .75)},
    'y': {'i': 1, 'tickdir': 0, 'juggled': (0, 1, 2),
          'color': (0.20, 0.90, 0.90, 0.25)},
    'z': {'i': 2, 'tickdir': 0, 'juggled': (0, 2, 1),
          'color': (0.925, 0.125, 0.90, 0.25)},}

class custom_XAxis(axis3d.Axis):
    _AXINFO = custom_AXINFO

class custom_YAxis(axis3d.Axis):
    _AXINFO = custom_AXINFO

class custom_ZAxis(axis3d.Axis):
    _AXINFO = custom_AXINFO

class custom_Axes3D(Axes3D):
    def _init_axis(self):
        '''Init 3D axes; overrides creation of regular X/Y axes'''
        self.w_xaxis = custom_XAxis('x', self.xy_viewLim.intervalx,
                                    self.xy_dataLim.intervalx, self)
        self.xaxis = self.w_xaxis
        self.w_yaxis = custom_YAxis('y', self.xy_viewLim.intervaly,
                            self.xy_dataLim.intervaly, self)
        self.yaxis = self.w_yaxis
        self.w_zaxis = custom_ZAxis('z', self.zz_viewLim.intervalx,
                            self.zz_dataLim.intervalx, self)
        self.zaxis = self.w_zaxis

        for ax in self.xaxis, self.yaxis, self.zaxis:
            ax.init3d()

# The rest of your code below, note the call to our new custom_Axes3D

points = (5*np.random.randn(3, 50)+np.tile(np.arange(1,51), (3, 1))).transpose()
fig = plt.figure(figsize = (10,10)) 
ax = custom_Axes3D(fig)

This is monkey-patching at it's worst, and should not be relied upon to work for later versions.

Fixing the facecolors was easier than the grid lines, as this requires an override of one of the __init__ methods, though it could be done with more work.

It does not seem difficult to expose this to the end user, and as such I can imagine that this may be fixed in later versions.