Let's apply the brand image color to the matplotlib colormap!

Introduction

What colors do you use when drawing graphs with matplotlib? Matplotlib comes with a beautiful colormap by default, I want to use different color variations for presentations, materials and graphs Have you ever wanted to unify the color image?

In this article, the brand color image to the matplotlib colormap I will challenge adaptation!

Class to use

Matplotlib has two types of classes for creating your own colormap.

matplotlib.colors.ListedColormap

Create by setting the color code that composes the colormap with list or array. The constituent colors of the created colormap change discretely.

[Usage example]

import matplotlib as mpl
color_list = ['red', 'green', 'blue']
cmap = mpl.colors.ListedColormap(color_list)

[Created color map]

matplotlib.colors.LinearSegmentedColormap

Create by setting the RGB information that makes up the colormap in the specified data format. The constituent colors of the created colormap change continuously.

[Usage example]

import matplotlib as mpl

#########################################################################
# 
#[Constituent color settings]
#
#The data of the constituent colors is,
# [
#  (x_0, y_0, z_0)
#  ...
#  (x_i, y_i, z_i)
# ]
#The RGB value is set to a value between 0 and 1.
#The constituent colors of colormap are z_0 -> y_1, z_1 -> y_2, ...It changes continuously in the order of.
#Red with the following settings-> green, green ->It means a change of blue.
#
#########################################################################

segment_data = {    
    'red':
        [
            (0.0,   0/255, 255/255),
            (0.5,   0/255,   0/255),
            (1.0,   0/255,   0/255),
        ],
    'green':
        [
            (0.0,   0/255,   0/255),
            (0.5, 128/255, 128/255),
            (1.0,   0/255,   0/255),
        ],
    'blue':
        [
            (0.0,   0/255,   0/255),
            (0.5,   0/255,   0/255),
            (1.0, 255/255,   0/255),
        ],
}
cmap = mpl.colors.LinearSegmentedColormap('colormap_name', segment_data)

With the above implementation, setting the constituent colors is not intuitive and a little difficult, but Using matplotlib.colors.LinearSegmentedColormap.from_list It can be created with a list of color codes.

#When creating with list
import matplotlib as mpl
color_list = ['red', 'green', 'blue']
cmap = mpl.colors.LinearSegmentedColormap.from_list('colormap_name', color_list)

[Created color map]

Implementation

When actually creating your own colormap, use the colormap created with matplotlib.cm.register_cmap. It is convenient to register.

import matplotlib.pyplot as plt
import matplotlib as mpl

def set_custom_colormap(name: str, color_list: list):
    
    """
    color_Convert list to colormap and register it in matplotlib with name.
Two patterns of discrete type and continuous type and four patterns of colormap are registered in reverse order.
    """

    cmap_dis = mpl.colors.ListedColormap(color_list)
    cmap_seq = mpl.colors.LinearSegmentedColormap.from_list(name, color_list)
    
    plt.register_cmap(name + '_dis', cmap_dis)
    plt.register_cmap(name + '_dis_r', cmap_dis.reversed())
    plt.register_cmap(name + '_seq', cmap_seq)
    plt.register_cmap(name + '_seq_r', cmap_seq.reversed())

    return

After that, you can check the color code of the color you want to set.

SENSY colormap

COLORMAP_SOURCE_DICT = {
    
    #Single color gradient
    'sensy_single_1': ['#FFFFFF', '#F4458C'],
    'sensy_single_2': ['#FFFFFF', '#FF6B9A'],
    'sensy_single_3': ['#FFFFFF', '#FA8EB5'],
    'sensy_single_4': ['#FFFFFF', '#A7C0FD'],
    'sensy_single_5': ['#FFFFFF', '#5073ED'],
    'sensy_single_6': ['#FFFFFF', '#4E6FF0'],
    'sensy_single_7': ['#FFFFFF', '#4B4C80'],
    'sensy_single_8': ['#FFFFFF', '#2B2C4B'],
    
    #Multicolor gradient
    'sensy': ['#FF6B9A', '#4E6FF0'],
    'sensy_diverge': ['#FF6B9A', '#FFFFFF', '#4E6FF0'],
    'sensy_accent': ['#5073ED', '#FA8EB5', '#A7C0FD', '#2B2C4B', '#F4458C', '#4B4C80'],

}

#colormap creation&Registration
for name, color_list in COLORMAP_SOURCE_DICT.items():
    set_custom_colormap(name, color_list)


#Initial setting###################

#Initial colormap
DEFAULT_COLORMAP = 'sensy_accent_dis'

#When using matplotlib
mpl.rcParams['image.cmap'] = DEFAULT_COLORMAP

#When using seaborn
import seaborn as sns
sns.set()
sns.set_palette(DEFAULT_COLORMAP)

[Color map image]

If the suffix of the colormap name is'seq', it is a continuous type, and if'dis' is a discrete type. If you add'_r' at the end, the order of colors will be reversed.

[Adaptation image]

Graph type	image
bar graph
Scatter plot
contour
Heat map

The data and code used for drawing the graph are based on the Official seaborn tutorial.

in conclusion

If I knew the color code I wanted to use, I could easily create my own colormap. If you feel unsatisfied or inconsistent with the color variations of the graphs you usually use, It may be fun to make a beautiful colormap by yourself.

Let's apply the brand image color to the matplotlib colormap!

Introduction

Class to use

Implementation

in conclusion

reference