Matplotlib

Every matplotlib question is really “which object owns this?”

Why half the examples online look nothing like the other half

Six lines, and five of them are about being readable

Backends, the blank-file trap, and getting a file worth sending

What the segment between two points is actually claiming

Do these two columns move together — and how much can you read into that

Sort them, turn them sideways, and never cut the baseline

What one column looks like before you average it

Locators say where, formatters say what — almost every tick problem is one of the two

Log axes, zero baselines, and why twin axes can prove anything

Three families, and picking the wrong one invents structure that is not there

Where a chart stops showing data and starts making a claim

Small multiples, and the array shape that breaks your loop

Draw the arrangement you want and get back Axes named after it

constrained layout, figure-level labels, and one legend for the lot

Decide how your charts look once, in one place

df.plot() is a shortcut into matplotlib, not an alternative to it

The ax=None convention, and sizing a figure for where it is going