Statistical plots with seaborn

Draw seaborn’s statistical plots — bars, boxes, distributions — onto a themed, on-brand chart.

How it works

seaborn draws onto a matplotlib Axes, and a managed render callback hands you exactly that — so seaborn works with no special integration. Call it with ax=ax:

import seaborn as sns

def render(ax):
    sns.barplot(data=rows, x="quarter", y="revenue", hue="region", ax=ax)

bs.Chart(render=render, grow=True)

When seaborn is installed and imported, Chart seeds its palette from the theme’s accent colors, so a categorical plot is on-brand and flips with light/dark — no palette= needed. Because seaborn’s plots are usually area-filled (bars, violins, KDE), the seeded colors are softened from the full accent saturation; tune that with seaborn_desat (0–1, default 0.75), or pass an explicit palette= to override entirely.

Note

Install the extra with pip install bootstack[viz-seaborn]. seaborn stays off the import path — Chart only seeds its palette when you have imported seaborn.

The same axes takes any seaborn plot — swap the one call for the chart you need:

sns.boxplot(data=rows, x="group", y="value", ax=ax)        # distributions
sns.violinplot(data=rows, x="group", y="value", ax=ax)
sns.heatmap(matrix, ax=ax)                                 # a matrix
sns.regplot(data=rows, x="x", y="y", ax=ax)                # a fit line

Example

# A small tidy dataset: revenue by quarter, split by region.
QUARTERS = ["Q1", "Q2", "Q3", "Q4"]
REGIONS = ["North", "South", "East"]
DATA = {"quarter": [], "region": [], "revenue": []}
for i, q in enumerate(QUARTERS):
    for j, r in enumerate(REGIONS):
        DATA["quarter"].append(q)
        DATA["region"].append(r)
        DATA["revenue"].append(20 + 8 * i + 5 * j)


def render(ax):
    """A grouped bar chart — seaborn picks up the seeded accent palette."""
    sns.barplot(data=DATA, x="quarter", y="revenue", hue="region", ax=ax)
    ax.set_xlabel("")
    ax.set_ylabel("revenue ($k)")


with bs.App(title="Statistical plots", size=(640, 460), padding=16, gap=12) as app:
    bs.Label("A seaborn bar chart, on-brand and themed", font="heading-md")
    bs.Chart(render=render, grow=True, horizontal="stretch")
    bs.Button("Toggle theme", on_click=bs.toggle_theme)

app.run()

When to use

Reach for seaborn when you want statistical summaries — grouped bars, box and violin plots, heatmaps, regressions — with minimal code. For plain line/bar/ scatter plots, matplotlib alone is enough (Plotting your data). seaborn plots are reactive too: drive them from a signal or data source exactly as in Live and data-driven charts. Palette and seaborn_desat details are on the Chart guide.