How do you choose bin width for a histogram?

medium

Answer

$(3.5{\sigma}\; / \;n(1 / 3), \;\mathrm{assumes}\;\mathrm{near} - \mathrm{normal})$
$Modern default in most libraries: Freedman-Diaconis.$
$Too few bins hide structure; too many create noisy jaggedness.$
$Best practice: try 2-3 bin widths, or use KDE (kernel density estimate) for a smooth alternative.$

Check yourself — multiple choice

One bin
Freedman-Diaconis $(2 \cdot \mathrm{IQR} / n(1 / 3))$ is robust default; too few bins hide structure, too many are noisy; KDE is smooth alternative
Always 10 bins
Bin width is irrelevant

Bin width: F-D robust default; try 2-3 or KDE for smooth.

#descriptive#eda

215 interview questions in this topic.