[FIXED] why ax.transData.transform([0.5, 0.5]) and ax.transAxes.transform([0.5, 0.5]) give the same result when ax = fig.add_subplot(2, 2, i+1)?

Issue

I was reading Data coordinates and Axes coordinates

fig = plt.figure()
for i, label in enumerate(('A', 'B', 'C', 'D')):
    ax = fig.add_subplot(2, 2, i+1)
    print(ax.transData.transform([0.5, 0.5]))
    print(ax.transAxes.transform([0.5, 0.5]))
    # why the two print the same result? 
    
    ax.text(0.1, 0.5, label, transform=ax.transAxes,
            fontsize=16, fontweight='bold', va='top')
    # transform=ax.transAxes replaced by transform=ax.transData the picture remains same 
plt.show()

I suppose that when ax created by fig.add_subplot(2, 2, i+1), these ax have different coordinate in their parent figure, so ax.transData should defer from ax.transAxes when passed the same args, but the result perplexed me.

Solution

By default if the axes limits are not specified, data coordinates and axes coordinates will be the same and are normalized to the limits of 0 to 1 within the limits of the axes.

You can set the limits of x and y for your data coordinates by these code:

ax.set_xlim(min_x, max_x)
ax.set_ylim(min_y, max_y)

To see the effect, run this modified code:-

import matplotlib.pyplot as plt

fig = plt.figure()
for i, label in enumerate(('A', 'B', 'C', 'D')):
    ax = fig.add_subplot(2, 2, i+1)
    ax.set_xlim(0, 8) # data coordinates
    ax.set_ylim(0, 6)
    print(ax.transData.transform([0.5, 0.5]))
    print(ax.transAxes.transform([0.5, 0.5]))
    # the two now print different results

    ax.text(0.1, 0.5, label, transform=ax.transAxes,
            fontsize=16, fontweight='bold', va='top')

    ax.text(4.1, 3.5, label+"data", transform=ax.transData,
        fontsize=16, fontweight='bold', va='top')
plt.show()

Note that the texts A,B,C,D are now plotted at the new data coordinates as labeled on the borders. But these coordinates are actually equivalent to the normalized (0 to 1) of the axes coordinates.

Answered By - swatchai