Seaborn · Function

color_palette

Build a Seaborn palette from a named palette, colormap, palette specification, or explicit colors. It can return discrete colors or a continuous colormap.

How it works

color_palette() accepts a named Seaborn or Matplotlib palette, a supported palette family specification, or a sequence of Matplotlib-compatible colors. It returns a color-palette object by default; setting as_cmap=True instead produces a continuous colormap. The number of colors and saturation can be adjusted with n_colors and desat.

Syntax

seaborn.color_palette(palette=None, n_colors=None, desat=None, as_cmap=False)

Parameters

Name	Type	Description
`palette`	`string, sequence, or None`	Palette definition: a Seaborn palette name, Matplotlib colormap name, supported palette specification, or sequence of Matplotlib-compatible colors.
`n_colors`	`int or None`	Number of colors in the palette. If omitted, the default depends on the palette specification.
`desat`	`float or None`	Proportion by which to desaturate each color.
`as_cmap`	`bool`	If true, return a continuous colormap instead of a discrete palette.

Returns

A list-like color palette, or a continuous colormap when as_cmap=True.

Key points

Use the function for either a discrete color collection or a continuous colormap.
Several built-in qualitative Seaborn palettes can be selected directly by name.
Matplotlib colormap names are valid palette specifications.
The HUSL and HLS hue systems are available through the palette argument.
You can provide explicit colors using formats supported by Matplotlib.
The discrete result wraps RGB tuples and provides palette-oriented utilities.
The returned palette supports hexadecimal conversion and rich display in HTML environments.

Common mistakes

n_colors has no effect when requesting a continuous colormap with as_cmap=True.
Requesting more colors than a finite palette provides repeats its colors.

Example: Create a discrete palette and a colormap

import seaborn as sns

palette = sns.color_palette("deep", n_colors=4)
print(palette.as_hex())

cmap = sns.color_palette("Spectral", as_cmap=True)
print(cmap(0.5))

Creates four colors from Seaborn’s named categorical palette and converts them to hexadecimal strings, then creates a continuous colormap from a ColorBrewer palette.

When to use it

Choose a named qualitative palette

Need consistent categorical colors for a plot.

Create a continuous colormap

Need a scalar-to-color mapping for a heatmap or continuous plot.

Supply explicit Matplotlib-compatible colors

Need exact brand or domain colors.

Adjust saturation with desat

Need a lighter or less saturated visual treatment.

Interview questions

QWhat kinds of values can palette accept?

The palette argument can be a Seaborn palette name, a Matplotlib colormap name, husl or hls, supported palette specifications such as ch:<cubehelix arguments>, or a sequence of Matplotlib-compatible colors. With palette=None, the function uses the current Matplotlib color cycle.

Follow-up: How would you obtain a continuous mapping rather than discrete colors? Pass as_cmap=True; the return value is then a continuous colormap.

QHow does n_colors behave?

n_colors controls the requested palette length when a discrete palette is returned. If it is omitted, the default depends on how palette is specified; named palettes generally default to six colors.

Follow-up: What happens if the requested length exceeds the palette size? The palette cycles through its available colors.

QWhat does color_palette() return?

The default return is a list-like Seaborn palette object containing RGB tuples and methods such as hexadecimal conversion. With as_cmap=True, the function returns a continuous colormap instead.

Follow-up: How can the discrete result be made easier to serialize for web use? Call its as_hex() method to obtain hexadecimal color strings.

set_palette
husl_palette
hls_palette
cubehelix_palette
dark_palette
light_palette
diverging_palette
blend_palette

Source: seaborn.color_palette. Checked 2026-10-15.