SeabornStyle#

Download this notebook from GitHub (right-click to download).

import seaborn as sns
import panel as pn
import matplotlib.font_manager

from matplotlib.figure import Figure

pn.extension(sizing_mode="stretch_width", template="fast")

Styling Seaborn for Panel#

In this example we will show how to style Seaborn charts with Panel for both the default and the dark theme.

Get or set the theme#

When we use the Fast templates the theme can be found in the session_args.

def get_theme():
    return pn.state.session_args.get("theme", [b'default'])[0].decode()

theme = get_theme()
theme

'default'

Select a nice accent color#

Below we create some functionality to cycle through a list of nice accent colors. You would probably just set the accent_color and color for your specific use case.

nice_accent_colors = [
    ("#00A170", "white"), # Mint
    ("#DAA520", "white"), # Golden Rod
    ("#F08080", "white"), # Light Coral
    ("#4099da", "white"), # Summery Sky
    ("#2F4F4F", "white"), # Dark Slate Grey
    ("#A01346", "white"), # Fast
]

def get_nice_accent_color():
    """Returns the 'next' nice accent color"""
    if not "color_index" in pn.state.cache:
        pn.state.cache["color_index"]=0
    elif pn.state.cache["color_index"]==len(nice_accent_colors)-1:
        pn.state.cache["color_index"]=0
    else:
        pn.state.cache["color_index"]+=1
    return nice_accent_colors[pn.state.cache["color_index"]]

accent_color, color = get_nice_accent_color()
pn.pane.Markdown(f"# Color: {accent_color}", background=accent_color, height=70, margin=0, style={"color": color, "padding": "10px"})

Seaborn: `set_theme`#

You can set the style, palette and font using sns.set_theme. See https://seaborn.pydata.org/generated/seaborn.set_theme.html.

def plot(style="white", palette="deep", font="sans-serif", theme="default", rc={}):
    sns.set_theme(style=style, palette=palette, font=font, rc=rc)
    
    fig = Figure(figsize=(12, 6))
    ax = fig.add_subplot(111)

    sns.barplot(x=["A", "B", "C"], y=[1, 3, 2], ax=ax)
    return fig

Lets’s add the Seaborn logo to the app

pn.Row(
    pn.layout.HSpacer(),
    pn.pane.SVG(
        "https://seaborn.pydata.org/_images/logo-tall-lightbg.svg",
        sizing_mode="fixed",
        width=210,
    ),
    pn.layout.HSpacer(),
).servable(target="sidebar")

We can use Select widgets to explore the style, palette and font arguments

STYLES = [
    "dark",
    "ticks",
    "white",
    "whitegrid",
]
PALETTES = [
    "bright",
    "colorblind",
    "dark",
    "deep",
    "hls",
    "husl",
    "muted",
    "pastel",
    f"dark:{accent_color}",
    f"light:{accent_color}",
]
if theme == "dark":
    PALETTE = f"light:{accent_color}"
else:
    PALETTE = f"dark:{accent_color}"

FONTS = sorted(set([f.name for f in matplotlib.font_manager.fontManager.ttflist]))

if theme=="dark":
    RC = {
        "axes.labelcolor": "white",
        "axes.facecolor": "black",
        "figure.facecolor": "black",
        "xtick.color": "white",
        "ytick.color": "white",
    }
else:
    RC = {}

style = pn.widgets.Select(name="Style", value="white", options=STYLES).servable(target="sidebar")
palette = pn.widgets.Select(name="Palette", value=PALETTE, options=PALETTES).servable(
    target="sidebar"
)
font = pn.widgets.Select(name="Font", value="Verdana", options=FONTS).servable(target="sidebar")
pn.Column(style, palette, font)

Lets define a plot

We can bind the plotfunction to the widgets using pn.bind

interactive_plot = pn.bind(plot, style=style, palette=palette, font=font, theme=theme, rc=RC)
pn.panel(interactive_plot, sizing_mode="scale_width").servable()

Style the app template#

pn.state.template.param.update(
    site="Panel", title="Styling Seaborn", header_background=accent_color, accent_base_color=accent_color, favicon="https://raw.githubusercontent.com/mwaskom/seaborn/master/doc/_static/favicon.ico",
)

Serve the app#

You can serve the app via panel serve SeabornStyle.ipynb and find it at http://localhost:5006/SeabornStyle. You should add the --autoreload flag while developing for hot reloading.

This web page was generated from a Jupyter notebook and not all interactivity will work on this website. Right click to download and run locally for full Python-backed interactivity.