Legend specifying 3d position in 3D axes matplotlib

Question

I have a scatter 3d plot using matplotlib.

What I'm trying to do is to position the legend inside the plot. I have read the documentation and it seems that is only possible to select predefined positions or only specify x and y coordinates.

Is there a way to position the legend specifying the 3 coordinates?

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
#more code
ax.legend(loc=(0.5,0.5,0.5), frameon=0)

The last line is what I thought might work but obviously is not working.

This is what I've now:

I'm trying to position the legend inside the axes, sort of like:

I reached that position by trial and error using ax.legend(loc=(0.15,0.65),frameon=0) because the legend doesn't move as the axes are rotated. The issue is that I'll be doing several plots thus I'm trying to avoid the trial and error approach.

Thanks.

Answer 1

To place the legend in a 3D plot using data coordinates, one may first get the projected coordinates of a point in 3D space using

mpl_toolkits.mplot3d.proj3d.proj_transform(x,y,z, ax.get_proj())

and provide those to the bbox_to_anchor argument of the legend. Than changing the bbox_transform to the data coordinate system produces the desired plot.

The following places the lower left corner of the legend at position (70,1000,80) in data coordinates.

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D, proj3d
import numpy as np

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')

theta = np.linspace(-4 * np.pi, 4 * np.pi, 100)
z = np.linspace(-2, 2, 100)
r = z**2 + 1
x = 25 * r * np.sin(theta)
y = 350* r * np.cos(theta)
ax.plot(x, y, 70*z, label='parametric curve')
ax.plot(x*.6, y*0.5, 70*z, label='parametric curve 2')

f = lambda x,y,z: proj3d.proj_transform(x,y,z, ax.get_proj())[:2]
ax.legend(loc="lower left", bbox_to_anchor=f(70,1000,80), 
          bbox_transform=ax.transData)

plt.show()