Using Custom Hex Colors and RGB Palettes in Matplotlib

TL;DR

Pass a hex string or an RGB tuple to any color argument:

ax.bar(categories, values, color='#2A9D8F')               # hex
ax.plot(x, y,              color=(0.165, 0.616, 0.557))   # RGB (0-1 scale)

Bar chart with five custom hex colors

Hex colors

Any CSS-style hex code works — three-char (#2AF) or six-char (#2A9D8F), with or without an alpha channel (#2A9D8F80 for 50% opacity).

import matplotlib.pyplot as plt

categories = ['Alpha', 'Beta', 'Gamma', 'Delta', 'Epsilon']
values     = [42, 67, 53, 81, 38]
hex_colors = ['#264653', '#2A9D8F', '#E9C46A', '#F4A261', '#E76F51']

fig, ax = plt.subplots(figsize=(7, 4))
ax.bar(categories, values, color=hex_colors)
plt.tight_layout()
plt.show()

You can pass a single hex string to color every bar the same, or a list to color each one individually.

RGB tuples

Matplotlib expects normalised RGB values in the range 0.0–1.0 (not 0–255). Divide your 8-bit values by 255 to convert:

# 8-bit values (from a design tool, for example)
r, g, b = 42, 157, 143
color = (r / 255, g / 255, b / 255)   # (0.165, 0.616, 0.557)

ax.plot(x, y, color=color)

Building a palette as a list of tuples

rgb_palette = [
    (0.149, 0.274, 0.325),   # dark teal
    (0.165, 0.616, 0.557),   # mid teal
    (0.953, 0.604, 0.239),   # amber
    (0.878, 0.275, 0.063),   # orange-red
]

for func, color, label in zip(functions, rgb_palette, labels):
    ax.plot(x, func, color=color, lw=2, label=label)

Line chart using a four-color custom RGB palette

Building a reusable palette dict

Keeping colors in a dict makes it easy to reference them by name across multiple charts:

palette = {
    'charcoal':      '#264653',
    'persian_green': '#2A9D8F',
    'sandy_yellow':  '#E9C46A',
    'sandy_brown':   '#F4A261',
    'burnt_sienna':  '#E76F51',
}

ax.bar(categories, values, color=[palette[k] for k in palette])

Visualising your palette as swatches

Before committing to a palette it helps to preview it. A quick strip of Rectangle patches does the job:

import matplotlib.pyplot as plt
import matplotlib.patches as mpatches

palette = {
    '#264653': 'Charcoal',
    '#2A9D8F': 'Persian Green',
    '#E9C46A': 'Sandy Yellow',
    '#F4A261': 'Sandy Brown',
    '#E76F51': 'Burnt Sienna',
}

fig, ax = plt.subplots(figsize=(7, 1.6))
ax.set_xlim(0, len(palette))
ax.set_ylim(0, 1)
ax.axis('off')

for i, (hex_val, name) in enumerate(palette.items()):
    ax.add_patch(plt.Rectangle((i, 0), 1, 0.7, color=hex_val))
    ax.text(i + 0.5, 0.78, hex_val, ha='center', va='bottom', fontsize=8.5)
    ax.text(i + 0.5, -0.08, name,    ha='center', va='top',    fontsize=7.5)

plt.tight_layout()
plt.show()

Color swatch strip showing the five custom palette colors

Adding transparency (alpha)

All color arguments accept an optional alpha keyword (0.0–1.0), or you can bake it into the hex code as a fourth byte:

ax.bar(categories, values, color='#2A9D8F', alpha=0.7)   # keyword
ax.bar(categories, values, color='#2A9D8FB3')             # hex with alpha byte

Where colors can be used

Any Matplotlib element that accepts a color argument works the same way — lines, bars, scatter points, patches, text, spines, and tick labels:

ax.plot(x, y, color='#E76F51')
ax.scatter(x, y, color='#2A9D8F', edgecolors='#264653')
ax.set_xlabel('x', color='#636e72')
ax.spines['bottom'].set_color('#264653')
ax.tick_params(colors='#636e72')

Python Charts

Python plotting and visualization demystified