Matplotlib colors

How to use colors in matplotlib

Every matplotlib function that draws something (plot, scatter, bar, hist, fill_between...) accepts a color argument (or c for short). The same color can be written in several equivalent ways, so pick the one that fits your workflow:

import matplotlib.pyplot as plt

fig, ax = plt.subplots()

ax.plot([1, 2, 3], color = "tomato")             # Named color
ax.plot([2, 3, 4], color = "r")                  # Single letter shorthand
ax.plot([3, 4, 5], color = "#1E90FF")            # HEX code
ax.plot([4, 5, 6], color = (0.2, 0.4, 0.6))      # RGB tuple (0-1)
ax.plot([5, 6, 7], color = "tab:orange")         # Tableau palette
ax.plot([6, 7, 8], color = "C3")                 # Fourth color of the default cycle
ax.plot([7, 8, 9], color = "xkcd:sky blue")      # XKCD color survey
ax.plot([8, 9, 10], color = "0.75")              # Gray level between 0 and 1

plt.show()

Named colors

Matplotlib recognizes the 148 CSS colors listed above by name, such as "tomato", "steelblue" or "darkorange". Names are case-insensitive and are the most readable option.

Single letter colors

The eight base colors have a one-letter shorthand: "b" (blue), "g" (green), "r" (red), "c" (cyan), "m" (magenta), "y" (yellow), "k" (black) and "w" (white). They can also be combined with markers and line styles in a format string, for example plt.plot(x, y, "ro--") draws red circles joined by a dashed line.

HEX colors

Any color can be passed as a HEX string, like "#FF6347". Add two extra digits to set the transparency: "#FF634780" is the same color with 50% opacity.

RGB and RGBA tuples

Matplotlib expects RGB values between 0 and 1, not between 0 and 255. If you have a color like rgb(255, 99, 71), divide each value by 255 or use the color converter above, which returns both scales. A fourth value adds transparency (RGBA): (1, 0.39, 0.28, 0.5).

color = (255 / 255, 99 / 255, 71 / 255)          # rgb(255, 99, 71) in matplotlib

Tableau colors and the default color cycle

The Tableau palette ("tab:blue", "tab:orange", "tab:green"...) contains the ten colors that matplotlib uses by default. You can also refer to them by their position in the current color cycle with "C0" to "C9", which is handy to reuse the color of a previous series.

XKCD colors

Matplotlib also includes the 949 colors of the XKCD color survey. Use them with the xkcd: prefix, for example "xkcd:sky blue" or "xkcd:dusty rose".

Transparency

Besides RGBA tuples and 8-digit HEX codes, most functions have an alpha argument between 0 (transparent) and 1 (opaque): plt.scatter(x, y, color = "steelblue", alpha = 0.5).

Change the default colors

To use your own colors for every series without specifying them one by one, change the color cycle with rcParams. All the plots created afterwards will use these colors in order.

import matplotlib.pyplot as plt
from cycler import cycler

plt.rcParams["axes.prop_cycle"] = cycler(color = ["#264653", "#2A9D8F", "#E9C46A", "#F4A261", "#E76F51"])

If you prefer a complete look (background, grid and colors) use one of the matplotlib styles, and if you need to change the panel color see how to set the background color in matplotlib.

Get the list of colors in Python

The matplotlib.colors module stores all the named colors in dictionaries, so you can print them or convert them between formats:

import matplotlib.colors as mcolors

mcolors.BASE_COLORS      # 8 single letter colors
mcolors.TABLEAU_COLORS   # 10 Tableau colors
mcolors.CSS4_COLORS      # 148 CSS colors
mcolors.XKCD_COLORS      # 949 XKCD colors

mcolors.to_hex("tomato")          # '#ff6347'
mcolors.to_rgb("#FF6347")         # (1.0, 0.388, 0.278)
mcolors.to_rgba("tomato", 0.5)    # (1.0, 0.388, 0.278, 0.5)

Color scales and palettes

The colors on this page are single, fixed colors. If you need to map a numeric or categorical variable to color you need a color scale instead:

• Matplotlib colormaps: viridis, plasma, coolwarm, tab10 and the rest of colormaps for the cmap argument.
• Seaborn color palettes: palettes for the palette argument of seaborn.
• Plotly color scales: sequential, diverging and qualitative scales for plotly.

Seaborn accepts exactly the same color formats as matplotlib, and plotly accepts CSS color names, HEX codes and "rgb(255, 99, 71)" strings, so the colors on this page can be used with any of the three libraries.