3

In matplotlib 3d-plotting plot_surface even when the x, y and z axis limits are set to be >=0 the negative z-portion of the surface is still getting plotted. The same does not happen for 2d plots- if out of 20 data points provided as input for plotting, 10 fall outside the axis limits, say first-quadrant, they simply do not get displayed in the plot.

Please see the below code and corresponding plot as an evidence (code run in Jupyter notebook) -

import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits import mplot3d
from matplotlib import cm
%matplotlib notebook

x = np.linspace(0,1.5,100)
y = np.linspace(0,1.5,100)
X,Y = np.meshgrid(x,y)
Z = 1-X-Y
fig = plt.figure()
ax = fig.add_subplot(projection='3d')
surf = ax.plot_surface(X,Y,Z,cmap=cm.coolwarm)
ax.set_xlim(0,1.5)
ax.set_ylim(0,1.5)
ax.set_zlim(0,3)
ax.set_xlabel('x')
ax.set_ylabel('y')
ax.set_zlabel('z')
fig.colorbar(surf, shrink=0.5, aspect=5)
fig.savefig('Question10Fig')

Above plot

The existence of the blue colour along with the colour-bar alongside makes it evident that z<0 values are present in the plot. Why is that so?

Anirban Chakraborty
  • 539
  • 1
  • 5
  • 15

2 Answers2

1

That seems to origin from

ax.set_zlim(0,3)

Removing this line, the figure looks good to me. That seems to be a problem in matplotlib (see also this post).

Tobias Windisch
  • 984
  • 5
  • 16
  • 2
    thanks for sharing the reference post. It gave me the idea to use masking and thereby figure out a possible issue with mplot3d/matplotlib. The issue is that scatter plot respects masking while surface plots don't. I am explaining it in detail [here](https://stackoverflow.com/questions/67519618/matplotlib-mplot3d-scatterplot-respects-masking-but-surfaceplot-does-not-why). As for the point of removing the line `ax.set_zlim(0,3)` , it does not the solve the aim of not plotting portions outside of first-quadrant. It just stops the anomaly from being manifested. – Anirban Chakraborty May 13 '21 at 12:59
  • @AnirbanChakraborty: Of course, removing `ax.set_zlim` does not solve your problem, but it seems to be the origin of the problem. – Tobias Windisch May 13 '21 at 13:07
0

Add following code:

z[z < 0] = np.nan

just using set_zlim is not enough.

Tomerikoo
  • 18,379
  • 16
  • 47
  • 61
Chenchow Tsu
  • 15
  • 3