pyplot keeps a current figure and a current axes in hidden global state: plt.plot() draws onto whichever figure is current, and plt.savefig() writes it. dataviz has no current anything. A chart is a Plot value that you build, pass around and export – two charts built side by side never interfere, and nothing is drawn until you export.

pyplot	dataviz
plt.figure(figsize=(8, 5))	.size_inches(8, 5) on the chart
plt.plot(x, y)	line(x, y)
plt.scatter(x, y)	scatter(x, y)
plt.title("t"), plt.xlabel, plt.ylabel	title=, x_title=, y_title=, or .labels(...)
plt.xlim(a, b)	.scale_x_domain(a, b)
plt.savefig("f.svg")	save(chart, "f.svg")
plt.show()	save() to a file and open it
plt.subplot(2, 2, i)	render_facets(plots, cols=2)
plt.rcParams[...] = ...	a Theme, passed as theme= or with .theme(...)
plt.style.use("dark_background")	a preset from dataviz.core.themes: dark(), minimal(), high_contrast(), print_safe()

Two consequences worth knowing:

• Order does not matter. Because the chart is a description, calling .scale_x_domain() before or after .labels() gives the same chart; there is no “draw, then adjust” sequence to get right.
• Styling is per chart. A Theme travels with its Plot instead of living in global state, so one program can export a light and a dark version of the same chart without resetting anything in between. See themes and styling.

For the chart-by-chart mapping, see Coming from matplotlib.