A python package that creates colourblind friendly plots.

Overview

Colourblind8 is a brand new plotting package implemented for matplotlib to optimize graphs into a format interpretable by people with colourblindness. Colourblindness, also known as colour vision deficiency, is a condition that affects individuals how they perceive colours visually. According to colourblindawareness.org, colourblindness affects 1 in every 12 men and 1 in every 200 women globally. This package is inspired by the fact that people without knowledge about this condition don't know how to make their graphs accessible. Our objective is to enhance data visualization by implementing proper colours so that the graphs are perceived correctly by all readers as intended.

Scope

At this stage, Colourblind8 will focus on developing settings for the three most prevalent colourblind perspectives - protanopia, deuteranopia and tritanopia (see Appendix for more information on the types of colourblindness).

In this package, three functions are included to make matplotlibplots more interpretable for people with protanopia, deuteranopia and tritanopia. Within each function, users can select one of our customized colour palettes via the palette argument according to their preference: deutera for deuteranopia,prota for protanopia ,tritafor tritanopia. The following are our main functions:

• plot_lines(x, y, alpha = 1.0, labels = None, palette = None, title = None, x_lab = None, y_lab = None, legend_title = None):

  • This function implements a method that creates line graphs by initializing the colour of lines, the face colour of the chart, the transparency of objects, and the font size for title, legend and axes.

  Inputs: x variable, y variable (a list of lists), transparency, labels of each y variable, palette colour, title, x labels, y labels and legend title.
  Output: a matplotlib line graph object.

• plot_scatter(x, y, alpha = 1.0, labels = None, palette = None, title = None, x_lab = None, y_lab = None, legend_title = None):

  • This function implements a method that creates scatterplots by initializing the colour of data points, the face colour of the chart, the transparency of objects, and the font size for title, legend and axes.

  Inputs: x variable, y variable (a list of lists), transparency, labels of each y variable, palette colour, title, x labels, y labels and legend title.
  Output: a matplotlib scatterplot object.

• plot_histogram(y, alpha = 1.0, bins = None, labels = None, palette = None ,title = None, x_lab = None, legend_title = None):

  • This functions implements a method that enhances histograms through modifying the colour of histogram layers, the face colour of the chart, the transparency of objects,bin size and the font size for title, legend and axes.

  Inputs: y variable (a list of lists), transparency, bin size, labels of each y variable, palette colour, title, x labels, y labels and legend title.
  Output: a matplotlib histogram object.

For more details, please use the following commands:

help(Colourblind8) for class docstring

help(Colourblind8.modulename) for module docstring

To install Colourblind8 via pip, please input the following into the Terminal:

pip install git+https://github.com/UBC-MDS/Colourblind8.git

To upgrade to the latest version, use:

pip install -- upgrade git+https://github.com/UBC-MDS/Colourblind8.git

Please note, for a more clear execution of our package, Varada Kolhatkar (UBC MDS instructor) suggested an additional document (jupyter notebook) to demonstrate how to use our package with a test dataset, along with our tests being passed. She found this a more beneficial method of visually comparing the output with the test cases.

The demonstration can be found in the example folder of this repository named sample_usage.ipynb.

1. Since Colourblind8 is built on matplotlib, it is necessary to import matplotlib, as well as Colourblind8. Don't forget to initialize Colourblind8 before using.

import matplotlib.pyplot as plt
from colourblind8.colourblind8 import Colourblind8()

2. Select one of the available Colourblind8 graphs and one of our customized colour palettes for data visualization. We picked plot_lines() and palette = "trita" as an example:

cb = Colourblind8()
cb.plot_lines(x, y, alpha = 1.0, labels = None, palette = "trita", title = None, x_lab = None, y_lab = None, legend_title = None)

Three colour palettes are available : deutera,prota,trita

data:

x=[1,2,3,4]

y_list=[[1,2,3,4],[3,5,7,9],[4,7,5,9],[1,4,6,8],[4,6,7,9],[1,3,4,8],[0,4,7,2],[-1,4,7,-7],[9,7,5,3]]

cb=Colourblind8()
cb.plot_lines(x, y_list, alpha =1.0, labels =['a','b','c','d','e','f','g','h','i'], palette = 'trita', title = "Trita Line Example", x_lab = "X label",  y_lab = "Y label", legend_title = "Legend")

deutera	Prota	Trita

data:

N = 10
x = np.random.rand(N )
y_1 = np.random.rand(N)
y_2 = np.random.rand(N)
y_3 = np.random.rand(N)
y_4 = np.random.rand(N)
y_5 = np.random.rand(N)
y_6 = np.random.rand(N)
y_7 = np.random.rand(N)
y_8 = np.random.rand(N)
y_9 = np.random.rand(N)

y_list= []
for i in range(9):
    y = np.random.rand(N)
    y_list.append(y)

cb = Colourblind8()
cb.plot_scatter(x, y_list, alpha =1.0, labels =['a','b','c','d','e','f','g','h','i'], palette = 'deutera', title = "Deutera scatterplot Example", x_lab = "X label",  y_lab = "Y label", legend_title = "Legend" )

deutera	Prota	Trita

data:

N = 100
x = np.random.rand(N)
y = np.random.rand(N)
z = np.random.rand(N)
list_y = [x,y,z]

cb = Colourblind8()
cb.plot_histogram(y = list_y, palette= 'prota', x_lab=' X Label', title = 'Deutera Histogram Example', alpha = 0.5, bins =10, labels=['c', 'b', "c"], legend_title="legend")

deutera	Prota	Trita

As we are committed to support open source, we provide the complete Colourblind8 Control Flow Diagram for each function from our package for users who are interested in the backend workflow.

plot_lines() Control Flow Diagram plot_scatter() Control Flow Diagram plot_histogram() Control Flow Diagram

Branch Coverage Testing

(waiting for instructor's response on whether to use coverage.py or manual test)

As we are committed to support open source, we provide the complete Colourblind8 Control Flow Diagram for each function in our package for users who are interested in the backend workflow they can be found in the branch_coverage folder.

Branch Coverage Testing

In total, 6 tests were identified for our initial manual testing. We pass all 6 tests for each function and a screenshot in the img file named tests_results.png is available for your convenience.

To the best of our knowledge, there is currently no other colourblind specific package available for matplotlib that offers colourblind-friendly palettes or themes. Seaborn, in conjunction with matplotlib, offers a single default colour palette (source) named colorblind targeting readers with general colourblindness and does not accommodate different variants of colourblind conditions.

Colourblind8 will bridge the gap as it attempts to fit into the Python ecosystem by offering a convenient and effective way to give all matplotlib visualizations a theme and fully inclusive colour palette. A package that shows a similar contribution to the python ecosystem is daltonize. This package aids to creating all images colourblind friendly as opposed to Colourblind8 which will make specifically data visualizations, plots and graphs fully inclusive for all readers.

Protanopia

Protanopes are more likely to confuse:-

1. Black with many shades of red
2. Dark brown with dark green, dark orange and dark red
3. Some blues with some reds, purples and dark pinks
4. Mid-greens with some oranges

Deuteranopes

Deuteranopes are more likely to confuse:-

1. Mid-reds with mid-greens
2. Blue-greens with grey and mid-pinks
3. Bright greens with yellows
4. Pale pinks with light grey
5. Mid-reds with mid-brown
6. Light blues with lilac

Tritanopes

The most common colour confusions for tritanopes are light blues with greys, dark purples with black, mid-greens with blues and oranges with reds.

source:http://www.colourblindawareness.org/colour-blindness/types-of-colour-blindness/