Axes.autoscale has scale-inconsistent margins #16539

Bug report

Bug summary

Autoscaling axis behaviour is inconsistent for data on very small scales

Code for reproduction

import matplotlib.pyplot as plt
import numpy as np
x = np.repeat(np.arange(10), 10)
y = np.tile(np.arange(10), 10)

fig, axes = plt.subplots(1, 4, figsize=(16, 4))
axes[0].scatter(x, y)
axes[1].scatter(x*1e3, y*1e3)
axes[2].scatter(x*1e-3, y*1e-3)
axes[3].scatter(x*1e-6, y*1e-6)
plt.show()

Actual outcome

Expected outcome

The plot should scale in the same way regardless of absolute values of data range. E.g.

Matplotlib version

• Operating system: Arch Linux
• Matplotlib version: 3.0.3
• Matplotlib backend (print(matplotlib.get_backend())): module://ipykernel.pylab.backend_inline
• Python version: 3.8.1
• Jupyter version (if applicable): jupyter-client==5.3.4 jupyter-console==6.0.0 jupyter-core==4.6.1

Installed with pacman on arch linux (installs from source.)