Plotting overview

Every plot is a top-level function bv.plot_<type>(...) that returns (fig, ax). They all share the same conventions:

• Inputs are numpy arrays, or column names when you pass data= (see Data input).
• Styling comes from an optional spec= (the spec system).
• Any extra keyword is an override routed to the backend (color=, linewidth=, alpha=, label=, …).
• hue= / group= split one call into per-category series (Grouping).

The full set

Function	Draws	Page
plot_line	connected line	Basic
plot_scatter	markers	Basic
plot_step	step / staircase	Basic
plot_bar	bars (grouped/stacked)	Basic
plot_errorbar	points with error bars	Basic
plot_fill_between	shaded band	Basic
plot_violin	violin distributions	Distributions
plot_hexbin	2D density hexbins	Distributions
plot_pie	pie chart	Distributions
plot_image	heatmap / image + colorbar	2D & images
plot_vertical	vertical reference lines	Annotations
plot_horizontal	horizontal reference lines	Annotations
plot_pval	significance bracket	Annotations

Higher-level composed figures (raincloud, lollipop, …) live in Composite plots.

Channels: what each plot expects

behaviz validates inputs declaratively. Each plot declares channels — named, typed inputs (e.g. x is a vector, ys is a list of vectors, err must match y's length). Pass the wrong shape and you get a clear BehavizDataError naming the channel and the fix, not a backend traceback. The required channels for each plot are listed on its page.