matplotlib surface plot hides scatter points which should be in front

Question

Yet another question about matplotlib 3d surfaces... I have code which adds a scatter point to a matplotlib surface graph.

The problem that I have is that the point always appears behind the surface, regardless of which angle you view it from.

If I cobble an (admittedly ugly) version using 3 short lines to mark the same point, it is visible.

I have turned off the depthshade function, so it isn't this. Can anybody explain what is going on and how I can correct it? Here is a simplified version of the code:

import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import numpy as np

df = pd.DataFrame({10: {10: 1,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   15: {10: 4,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   20: {10: 6,15: 3,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   25: {10: 7,15: 5,20: 3,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   30: {10: 9,15: 6,20: 4,25: 3,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   35: {10: 10,15: 7,20: 5,25: 4,30: 2,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   40: {10: 11,15: 8,20: 6,25: 4,30: 3,35: 2,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   45: {10: 12,15: 9,20: 7,25: 5,30: 4,35: 3,40: 2,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   50: {10: 13,15: 9,20: 7,25: 6,30: 5,35: 4,40: 3,45: 2,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   55: {10: 14,15: 10,20: 8,25: 7,30: 5,35: 4,40: 3,45: 3,50: 2,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   60: {10: 15,15: 11,20: 9,25: 7,30: 6,35: 5,40: 4,45: 3,50: 3,55: 2,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   65: {10: 16,15: 12,20: 9,25: 8,30: 6,35: 5,40: 5,45: 4,50: 3,55: 2,60: 2,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   70: {10: 17,15: 12,20: 10,25: 8,30: 7,35: 6,40: 5,45: 4,50: 4,55: 3,60: 2,65: 2,70: 1,75: 1,80: 1,85: 1,90: 1},
                   75: {10: 18,15: 13,20: 10,25: 9,30: 7,35: 6,40: 5,45: 5,50: 4,55: 3,60: 3,65: 2,70: 2,75: 1,80: 1,85: 1,90: 1},
                   80: {10: 19,15: 14,20: 11,25: 9,30: 8,35: 7,40: 6,45: 5,50: 4,55: 4,60: 3,65: 3,70: 2,75: 2,80: 1,85: 1,90: 1},
                   85: {10: 21,15: 14,20: 11,25: 10,30: 8,35: 7,40: 6,45: 6,50: 5,55: 4,60: 4,65: 3,70: 3,75: 2,80: 2,85: 1,90: 1},
                   90: {10: 23,15: 15,20: 12,25: 10,30: 9,35: 8,40: 7,45: 6,50: 5,55: 5,60: 4,65: 3,70: 3,75: 3,80: 2,85: 2,90: 1}})




xv, yv = np.meshgrid(df.index, df.columns)
ma = np.nanmax(df.values)
norm = matplotlib.colors.Normalize(vmin = 0, vmax = ma, clip = True)

fig = plt.figure(1)
ax = Axes3D(fig)
surf = ax.plot_surface(yv,xv,df, cmap='viridis_r', linewidth=0.3,
                       alpha = 0.8, edgecolor = 'k', norm=norm)
ax.scatter(25,35,4, c='k', depthshade=False, alpha = 1, s=100)

fig = plt.figure(2)
ax = Axes3D(fig)
surf = ax.plot_surface(yv,xv,df, cmap='viridis_r', linewidth=0.3,
                       alpha = 0.8, edgecolor = 'k', norm=norm)
line1_x = [25,25]
line1_y = [35,35]
line1_z = [3,5]

line2_x = [25,25]
line2_y = [33,37]
line2_z = [4,4]

line3_x = [23,27]
line3_y = [35,35]
line3_z = [4,4]

ax.plot(line1_x, line1_y, line1_z, alpha = 1, linewidth = 1, color='k')
ax.plot(line2_x, line2_y, line2_z, alpha = 1, linewidth = 1, color='k')
ax.plot(line3_x, line3_y, line3_z, alpha = 1, linewidth = 1, color='k')
plt.show()

This is probably this known [issue with 3D projections.](https://matplotlib.org/mpl_toolkits/mplot3d/faq.html#my-3d-plot-doesn-t-look-right-at-certain-viewing-angles) — Mr. T, Jul 09 '18 at 08:49
The matplotlib documentation suggests MayaVi, so I guess not within matplotlib. — Mr. T, Jul 09 '18 at 08:50

score 8 · Answer 1 · answered Mar 10 '22 at 19:38

A useful workaround is to use the option computed_zorder=False (added in Feb 2021, see doc), and to plot the different elements in the desired order. The only caveat is that it requires knowing which points are below the surface, and which points are above.

import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import numpy as np

df = pd.DataFrame({10: {10: 1,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   15: {10: 4,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   20: {10: 6,15: 3,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   25: {10: 7,15: 5,20: 3,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   30: {10: 9,15: 6,20: 4,25: 3,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   35: {10: 10,15: 7,20: 5,25: 4,30: 2,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   40: {10: 11,15: 8,20: 6,25: 4,30: 3,35: 2,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   45: {10: 12,15: 9,20: 7,25: 5,30: 4,35: 3,40: 2,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   50: {10: 13,15: 9,20: 7,25: 6,30: 5,35: 4,40: 3,45: 2,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   55: {10: 14,15: 10,20: 8,25: 7,30: 5,35: 4,40: 3,45: 3,50: 2,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   60: {10: 15,15: 11,20: 9,25: 7,30: 6,35: 5,40: 4,45: 3,50: 3,55: 2,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   65: {10: 16,15: 12,20: 9,25: 8,30: 6,35: 5,40: 5,45: 4,50: 3,55: 2,60: 2,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   70: {10: 17,15: 12,20: 10,25: 8,30: 7,35: 6,40: 5,45: 4,50: 4,55: 3,60: 2,65: 2,70: 1,75: 1,80: 1,85: 1,90: 1},
                   75: {10: 18,15: 13,20: 10,25: 9,30: 7,35: 6,40: 5,45: 5,50: 4,55: 3,60: 3,65: 2,70: 2,75: 1,80: 1,85: 1,90: 1},
                   80: {10: 19,15: 14,20: 11,25: 9,30: 8,35: 7,40: 6,45: 5,50: 4,55: 4,60: 3,65: 3,70: 2,75: 2,80: 1,85: 1,90: 1},
                   85: {10: 21,15: 14,20: 11,25: 10,30: 8,35: 7,40: 6,45: 6,50: 5,55: 4,60: 4,65: 3,70: 3,75: 2,80: 2,85: 1,90: 1},
                   90: {10: 23,15: 15,20: 12,25: 10,30: 9,35: 8,40: 7,45: 6,50: 5,55: 5,60: 4,65: 3,70: 3,75: 3,80: 2,85: 2,90: 1}})




xv, yv = np.meshgrid(df.index, df.columns)
ma = np.nanmax(df.values)
norm = matplotlib.colors.Normalize(vmin = 0, vmax = ma, clip = True)

fig = plt.figure(1)
ax = Axes3D(fig, computed_zorder=False)

ax.scatter(10,70,4, c='k', depthshade=False, alpha = 1, s=100)
surf = ax.plot_surface(yv,xv,df, cmap='viridis_r', linewidth=0.3,
                       alpha = 0.8, edgecolor = 'k', norm=norm)
ax.scatter(25,35,4, c='k', depthshade=False, alpha = 1, s=100)

plt.show()

When I tried Axes3D I was getting a warning. It works however with plt.axes: "ax = plt.axes(projection ='3d', computed_zorder=False)". It took me a long time to find this solution. Thanks a lot! — Rational-IM, Apr 23 '22 at 23:38
This worked insanely well, thank you. Very clean, just one liner. — Sh.A, Nov 19 '22 at 17:43

score 5 · Accepted Answer · answered Jul 09 '18 at 10:23

OK, so as per the comment by Mr T above, there doesn't seem to be a direct method of dealing with this. There is however, a workaround for what I'm trying to do (highlight specific points on the surface). Using the matplotlib.patches and mpl_toolkits.mplot3d.art3d modules, it is possible to plot a circle on the graph at the appropriate point, and this appears to be unaffected by the same issue.

The modified code is:

import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D, art3d
from matplotlib.patches import Circle
import numpy as np

df = pd.DataFrame({10: {10: 1,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   15: {10: 4,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   20: {10: 6,15: 3,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   25: {10: 7,15: 5,20: 3,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   30: {10: 9,15: 6,20: 4,25: 3,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   35: {10: 10,15: 7,20: 5,25: 4,30: 2,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   40: {10: 11,15: 8,20: 6,25: 4,30: 3,35: 2,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   45: {10: 12,15: 9,20: 7,25: 5,30: 4,35: 3,40: 2,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   50: {10: 13,15: 9,20: 7,25: 6,30: 5,35: 4,40: 3,45: 2,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   55: {10: 14,15: 10,20: 8,25: 7,30: 5,35: 4,40: 3,45: 3,50: 2,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   60: {10: 15,15: 11,20: 9,25: 7,30: 6,35: 5,40: 4,45: 3,50: 3,55: 2,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   65: {10: 16,15: 12,20: 9,25: 8,30: 6,35: 5,40: 5,45: 4,50: 3,55: 2,60: 2,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   70: {10: 17,15: 12,20: 10,25: 8,30: 7,35: 6,40: 5,45: 4,50: 4,55: 3,60: 2,65: 2,70: 1,75: 1,80: 1,85: 1,90: 1},
                   75: {10: 18,15: 13,20: 10,25: 9,30: 7,35: 6,40: 5,45: 5,50: 4,55: 3,60: 3,65: 2,70: 2,75: 1,80: 1,85: 1,90: 1},
                   80: {10: 19,15: 14,20: 11,25: 9,30: 8,35: 7,40: 6,45: 5,50: 4,55: 4,60: 3,65: 3,70: 2,75: 2,80: 1,85: 1,90: 1},
                   85: {10: 21,15: 14,20: 11,25: 10,30: 8,35: 7,40: 6,45: 6,50: 5,55: 4,60: 4,65: 3,70: 3,75: 2,80: 2,85: 1,90: 1},
                   90: {10: 23,15: 15,20: 12,25: 10,30: 9,35: 8,40: 7,45: 6,50: 5,55: 5,60: 4,65: 3,70: 3,75: 3,80: 2,85: 2,90: 1}})




xv, yv = np.meshgrid(df.index, df.columns)
ma = np.nanmax(df.values)
norm = matplotlib.colors.Normalize(vmin = 0, vmax = ma, clip = True)

fig = plt.figure(1)
ax = Axes3D(fig)
surf = ax.plot_surface(yv,xv,df, cmap='viridis_r', linewidth=0.3,
                       alpha = 0.8, edgecolor = 'k', norm=norm)

p = Circle((25, 35), 3, ec='k', fc="none")
ax.add_patch(p)
art3d.pathpatch_2d_to_3d(p, z=4, zdir="z")

plt.show()

Thanks for the workaround. I'm facing the same issue in 2020, really wish there's a better solution — Logan Yang, Mar 23 '20 at 21:16

Z Li · Answer 3 · 2020-12-07T22:00:02.790

Run into this problem in 2020 and do not want to switch to another package. This solution is a modification of Will's answer above. Basically draw the circle in three axis to make it more like a dot. Also use ellipse to adjust for axis ratios. Works better if you set the radius smaller and choose a face color:

   def add_point(ax, x, y, z, fc = None, ec = None, radius = 0.005):
       xy_len, z_len = ax.get_figure().get_size_inches()
       axis_length = [x[1] - x[0] for x in [ax.get_xbound(), ax.get_ybound(), ax.get_zbound()]]
       axis_rotation =  {'z': ((x, y, z), axis_length[1]/axis_length[0]),
                         'y': ((x, z, y), axis_length[2]/axis_length[0]*xy_len/z_len),
                         'x': ((y, z, x), axis_length[2]/axis_length[1]*xy_len/z_len)}
       for a, ((x0, y0, z0), ratio) in axis_rotation.items():
           p = Ellipse((x0, y0), width = radius, height = radius*ratio, fc=fc, ec=ec)
           ax.add_patch(p)
           art3d.pathpatch_2d_to_3d(p, z=z0, zdir=a)

where radius is the radius of the "circle", fc is the face color, ec is the edge color.

The modified code:

import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D, art3d
from matplotlib.patches import Circle, Ellipse
import numpy as np

df = pd.DataFrame({10: {10: 1,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   15: {10: 4,15: 1,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   20: {10: 6,15: 3,20: 1,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   25: {10: 7,15: 5,20: 3,25: 1,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   30: {10: 9,15: 6,20: 4,25: 3,30: 1,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   35: {10: 10,15: 7,20: 5,25: 4,30: 2,35: 1,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   40: {10: 11,15: 8,20: 6,25: 4,30: 3,35: 2,40: 1,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   45: {10: 12,15: 9,20: 7,25: 5,30: 4,35: 3,40: 2,45: 1,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   50: {10: 13,15: 9,20: 7,25: 6,30: 5,35: 4,40: 3,45: 2,50: 1,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   55: {10: 14,15: 10,20: 8,25: 7,30: 5,35: 4,40: 3,45: 3,50: 2,55: 1,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   60: {10: 15,15: 11,20: 9,25: 7,30: 6,35: 5,40: 4,45: 3,50: 3,55: 2,60: 1,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   65: {10: 16,15: 12,20: 9,25: 8,30: 6,35: 5,40: 5,45: 4,50: 3,55: 2,60: 2,65: 1,70: 1,75: 1,80: 1,85: 1,90: 1},
                   70: {10: 17,15: 12,20: 10,25: 8,30: 7,35: 6,40: 5,45: 4,50: 4,55: 3,60: 2,65: 2,70: 1,75: 1,80: 1,85: 1,90: 1},
                   75: {10: 18,15: 13,20: 10,25: 9,30: 7,35: 6,40: 5,45: 5,50: 4,55: 3,60: 3,65: 2,70: 2,75: 1,80: 1,85: 1,90: 1},
                   80: {10: 19,15: 14,20: 11,25: 9,30: 8,35: 7,40: 6,45: 5,50: 4,55: 4,60: 3,65: 3,70: 2,75: 2,80: 1,85: 1,90: 1},
                   85: {10: 21,15: 14,20: 11,25: 10,30: 8,35: 7,40: 6,45: 6,50: 5,55: 4,60: 4,65: 3,70: 3,75: 2,80: 2,85: 1,90: 1},
                   90: {10: 23,15: 15,20: 12,25: 10,30: 9,35: 8,40: 7,45: 6,50: 5,55: 5,60: 4,65: 3,70: 3,75: 3,80: 2,85: 2,90: 1}})




xv, yv = np.meshgrid(df.index, df.columns)
ma = np.nanmax(df.values)
norm = matplotlib.colors.Normalize(vmin = 0, vmax = ma, clip = True)

fig = plt.figure(1)
ax = Axes3D(fig)
surf = ax.plot_surface(yv,xv,df, cmap='viridis_r', linewidth=0.3,
                       alpha = 0.8, edgecolor = 'k', norm=norm)

add_point(ax, 25, 35, 0, radius=1)
add_point(ax, 25, 35, 2, radius=2)
add_point(ax, 25, 35, 4, radius=3)

plt.show()

score 0 · Answer 4 · answered May 28 '22 at 13:42

0

Another solution would be to plot the point as a spherical surface. Then you can just plot the 'point' on top of the surface.

answered May 28 '22 at 13:42

Carl Bouchard

11

1

Your answer could be improved with additional supporting information. Please [edit] to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers [in the help center](/help/how-to-answer). – Community May 29 '22 at 06:43

matplotlib surface plot hides scatter points which should be in front

4 Answers4

Linked