Understanding Matplotlib's quiver plotting

Question

I'm trying to understand how plt.quiver() works. My issue is as follows:

I plot a simple vector (1,1) as such:

import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure(2)
ax = fig.add_subplot(111)
ax.quiver(0,0, 1, 1, units = 'xy', scale = 1)
plt.xticks(range(-5,6))
plt.yticks(range(-5,6))
plt.grid()

I would expect the arrow to go from (0,0) to (1,1), but the result is slightly off from that:

Similarly, I try and plot an arrow for vector (0,3) and the resulting arrow seems to be for vector (0,3.5)...

My assumption is that this has something to do with the kwargs 'units', 'scale', 'angles', & 'scale_units'. I've read the docs on them but don't fully understand how they work. A sunday school explanation would be greatly appreciated!

Answer 1

If you adjust the aspect ratio of the figure to 1, the vectors are displayed to proper scale:

import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure()
ax = fig.add_subplot(111)
ax.quiver((0,0), (0,0), (1,0), (1,3), units = 'xy', scale = 1)
plt.axis('equal')
plt.xticks(range(-5,6))
plt.yticks(range(-5,6))
plt.grid()
plt.show()

Answer 2

You can try this code.

import matplotlib.pyplot as plt

fig = plt.figure(2)
ax = fig.add_subplot(111)
ax.quiver(0,0, 1, 1,angles='xy', scale_units='xy', scale = 1)
plt.xticks(range(-5,6))
plt.yticks(range(-5,6))
plt.grid()
plt.draw()
plt.show()

Just remember that the first two arguments of quiver are the x and y coordinates of the tail of the vector, the next two are the lengths of the vector along x and y direction respectively. angle='xy' makes the arrow point from tail of the vector to its tip.

You can find out more about matplotlib.quiver at http://matplotlib.org/1.3.1/api/pyplot_api.html#matplotlib.pyplot.quiver