DOC: Fix example for what's new imshow so it isn't cut off or crowded. - #24441

Suggest constrained layout and concise date formatter.

By "concise date formatter", did you mean something like this? TBH this seems confusing...

I prefer the concise locator version, but I'm biased...

What about something like this?

I'm not going to block on that, but I don't find having 00:00 next to the day all that useful.

Since this example has so many ticks so close together, I think this helps avoid confusion of the ticks being multiline entries e.g., Jan-10\n12:00 kind of looks like a single label because the axis is so dense, but it is actually two labels Jan-10 00:00 and Jan-10 12:00. If anything, there is now a bonus of providing some ways to tweak the output of ConciseDateFormatter.

Here's the updated page in the docs. This is also 3.7 #22230

If you simply do:

import matplotlib.pyplot as plt
import numpy as np
plt.rcParams['date.converter'] = 'concise'

fig, ax = plt.subplots(layout='constrained')
date_first = np.datetime64('2020-01-01', 'D')
date_last = np.datetime64('2020-01-11', 'D')
arr = [[i+j for i in range(10)] for j in range(10)]

ax.imshow(arr, origin='lower', extent=[1, 11, date_first, date_last])

Without any monkeying with ticks or formatters, you get the pretty straight forward:

I would even leave out plt.rcParams['date.converter'] = 'concise'. IMHO this is not making it better here. Having only the numbers 02..11 on the y-axis (and a bit weird 'Jan') makes it less clear that the extent has been set to dates. So better without it:

Optional: I also switched the x-extent from [1, 11], to [0, 10] because it looks slightly better to have ticks on the boundaries. But that's just a minor stylistic thing.

I agree with @timhoffm – I will update the example.