CODEWITHSUNDEEP
pythonmatplotlibplot
john morrison
john morrison

Reputation: 187

Python doesn't let me change the number of axis ticks when using subplots

I have code which looks like this:

plt.figure()

plt.subplot(211)
plt.plot(timestamps_SW, np.random.randn(len(testLabelsSW_ds)), label='truth')
plt.plot(timestamps_SW, np.random.randn(len(testLabelsSW_ds)), label='pred.')
plt.locator_params(axis='y', nbins=5)
plt.legend()

plt.subplot(212)
plt.plot(timestamps_SE,np.random.randn(len(testLabelsSE_ds)), label='truth')
plt.plot(timestamps_SE,np.random.randn(len(testLabelsSE_ds)), label='pred.')
plt.locator_params(axis='y', nbins=5)
plt.legend()

The resulting plot looks as such:

[1]: https://i.sstatic.net/JsYKw.png

Clearly each y-axis doesn't have 5 ticks. How do I fix?

Upvotes: 1

Views: 64

Answers (1)

ImportanceOfBeingErnest
ImportanceOfBeingErnest

Reputation: 339052

Using plt.locator_params(..., nbins=5) you ask the default locator to use 5 bins. The default locator is an AutoLocator. It's a subclass of the MaxNLocator. MaxN means, it will try to find a maximum of N nice locations, where N is equal to nbins + 1. "Nice" locations means, e.g. something like 1.0, 0.25 etc are considered "nice", while something like 0.761 is of course not so nice.

The two constraints (a) "nice" and (b) "N locations" are of course seldomly fulfillable at the same time. Therefore "nice" has priority here.

If you leave out the constraint (a), you can get N locations simply by placing the ticks manually, 

plt.yticks(np.linspace(*plt.ylim(), 5))

Upvotes: 1

Related Questions