Describe › Standalone plots

Raincloud plot

A raincloud plot combines a half-violin density, a box plot and every jittered raw observation per group, showing distribution shape, summary and individual points in one figure.

What is Raincloud plot?

Bar charts hide distributions; box plots hide multimodality; violins hide the n. The raincloud stacks all three honest views: the density silhouette, the median/IQR box, and each raw observation jittered beside them.

Density uses a Gaussian kernel with Silverman's bandwidth, same engine as the violin plot; jitter is deterministic so redraws are stable.

When should I use Raincloud plot?

  • Comparing distributions across groups in presentations/papers — the current best-practice descriptive plot.
  • Any place you'd use a box plot but the audience should also see sample size and shape.

What data does it need?

One or more numeric columns, or one numeric column + a grouping column.

What does it report?

One cloud + box + rain triplet per group on a shared axis.

What does it assume?

  • Purely descriptive — none. With < ~10 points per group the density silhouette is unreliable; read the raw points instead.

How do I interpret the result?

Look for shape disagreements the box plot can't show: bimodality, skew, outlier clusters in the rain.

See also

References

  • Allen, Poggiali, Whitaker, Marshall & Kievit (2019). Raincloud plots: a multi-platform tool for robust data visualization. Wellcome Open Res 4.