matplotlib 3d plot with changing labels

Question

So I have a 3d live-updating graph! it only shows one point at a time so I can easily track the motion of the point! But here is the problem:

No matter what I seem to do, the point is always placed in the center of the graph and the tick marks on the axis change in order to do that. This makes my life very difficult because I don't see the motion on the point. Here is my code:

from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
from pylab import *
import time
import pandas as pd
import pickle
def pickleLoad(pickleFile):
    pkl_file = open(pickleFile, 'rb')
    data = pickle.load(pkl_file)
    pkl_file.close()
    return data
data = pickleLoad('/Users/ryansaxe/Desktop/kaggle_parkinsons/accelerometry/LILY_dataframe')
data = data.reset_index(drop=True)
df = data.ix[0:,['x.mean','y.mean','z.mean','time']]
ion()
fig = figure()
ax = fig.add_subplot(111, projection='3d')
count = 0
plotting = True
labels = range(-10,11)
ax.set_xlabel('X Label')
ax.set_ylabel('Y Label')
ax.set_zlabel('Z Label')
ax.set_yticklabels(labels)
ax.set_xticklabels(labels)
ax.set_zticklabels(labels)
lin = None
while plotting:
    df2 = df.ix[count]
    count += 1
    xs = df2['x.mean']
    ys = df2['y.mean']
    zs = df2['z.mean']
    t = df2['time']
    ax.set_title(t)
    if lin is not None:
        lin.remove()
    lin = ax.scatter(xs, ys, zs)

    draw()
    pause(0.01)
    if count > 100:
        plotting = False
ioff()
show()

here is an example of the data:

     x.mean    y.mean    z.mean                 time
0 -1.982905  3.395062  8.558263  2012-01-18 14:00:03
1  0.025276 -0.399172  7.404849  2012-01-18 14:00:04
2 -0.156906 -8.875595  1.925565  2012-01-18 14:00:05
3  2.643088 -8.307801  2.382624  2012-01-18 14:00:06
4  3.562265 -7.875230  2.312898  2012-01-18 14:00:07
5  4.441432 -7.907592  2.851774  2012-01-18 14:00:08
6  4.124187 -7.854146  2.727229  2012-01-18 14:00:09
7  4.199698 -8.135596  2.677706  2012-01-18 14:00:10
8  4.407856 -8.133449  2.214902  2012-01-18 14:00:11
9  4.096238 -8.453822  1.359692  2012-01-18 14:00:12

So what can I do to make it so that the tick marks are fixed so the point moves rather than the tick marks changing?

Answer 1

An Axes3D object (your ax variable) has the following methods: set_xlim, set_ylim, and set_zlim. You could use these to fix the limits of your axes.

Documentation:

• set_xlim
• set_xlim3d

Edit

Using set_xlim, etc, works for me. Here is my code:

#!python2

from mpl_toolkits.mplot3d import Axes3D
from pylab import *

data = [
    [-1.982905,  3.395062,  8.558263,  '2012-01-18 14:00:03'],
    [ 0.025276, -0.399172,  7.404849,  '2012-01-18 14:00:04'],
    [-0.156906, -8.875595,  1.925565,  '2012-01-18 14:00:05'],
    [ 2.643088, -8.307801,  2.382624,  '2012-01-18 14:00:06'],
    [3.562265, -7.875230,  2.312898,  '2012-01-18 14:00:07'],
    [4.441432, -7.907592,  2.851774,  '2012-01-18 14:00:08'],
    [4.124187, -7.854146,  2.727229,  '2012-01-18 14:00:09'],
    [4.199698, -8.135596,  2.677706,  '2012-01-18 14:00:10'],
    [4.407856, -8.133449,  2.214902,  '2012-01-18 14:00:11'],
    [4.096238, -8.453822,  1.359692,  '2012-01-18 14:00:12'],
]

ion()
fig = figure()
ax = fig.add_subplot(111, projection='3d')
ax.set_xlabel('X Label')
ax.set_ylabel('Y Label')
ax.set_zlabel('Z Label')
ax.set_xlim((-10, 11))
ax.set_ylim((-10, 11))
ax.set_zlim((-10, 11))

lin = None
for x, y, z, t in data:
    ax.set_title(t)
    if lin is not None:
        lin.remove()
    lin = ax.scatter(x, y, z)
    draw()
    pause(0.1)

ioff()
show()

Edit 2

You could have a look at switching off autoscaling of axes which is on by default. Maybe this is overriding the set_lim methods.

Documentation:

• autoscale
• autoscale_view

Answer 2

I ran into pretty much the exact same problem as you. I found one way to get it to work is to reset the axis limits each time you want to add a new point, so your loop would look something like:

while plotting:
    df2 = df.ix[count]
    count += 1
    xs = df2['x.mean']
    ys = df2['y.mean']
    zs = df2['z.mean']
    t = df2['time']
    ax.set_title(t)
    if lin is not None:
        lin.remove()
    lin = ax.scatter(xs, ys, zs)
    ax.set_xlim((-10, 11))
    ax.set_ylim((-10, 11))
    ax.set_zlim((-10, 11))

    draw()
    pause(0.01)
    if count > 100:
        plotting = False

I have a feeling resetting the limits everytime is a bad way to solve this, so I asked a question about it, and will hopefully find an answer. If all you want is to just see something working, then hopefully the limit changes within the loop will do the trick

Answer 3

You probably want to hold your axis-ranges constant. This can be done, for example, by axis().

See also these answers:
How to set axis range
y-axis limit
enforce axis range