Question

I am trying to put two scatter plots next to each other with a shared y axis, but the axis seems to get an odd scale. Without the shared axis the two plots look fine. I also noticed that the problem does not occur when using “plot” instead of “scatter”. Images are included below. Here is the code I am using.

#!/usr/bin/python

import matplotlib.pyplot as plt

fig = plt.figure(1)
for i in range(1,3):
    if i == 1:
        ax = fig.add_subplot(1,2,i)
    else:
        fig.add_subplot(1,2,i, sharey=ax)

    #plt.plot([5.0], [1],marker="*",color='tomato')
    plt.scatter([5.0], [1], s=20, color='tomato')

plt.show()

[I would include images but the site won’t let me as a newbie.] When I run the code above I see plots with a y axis that runs from 0.0000 to 0.0008 with a single point plotted at 0.0004. Without shared axes the y axis goes from 0.94 to 1.06 with a single point plotted at 1.00, as expected.

Can anyone tell me why? Is this a bug or a feature?

matplotlib: 0.99.1.2-3ubuntu on Ubuntu 10.04 LTS – the Lucid Lynx

Answer 1

I’ve no answer to the why question, but here’s how to get rid of it: In your code snippet, giving scatter three points to draw

plt.scatter([1.0,2,3], [1.1,2.2,2.9], s=20, color='tomato')

works for me (matplotlib 1.1.0 on Lucid).

I can only guess that scatter tries to be a bit smarter than plot with the axes limits,
but whatever it’s doing, it goes nuts for just a single point.