pySankey2

Uses matplotlib to create simple Sankey diagrams flowing from left to right.

A fork of a fork of pySankey.

Requirements

Requires python-tk (for python 2.7) or python3-tk (for python 3.x) install with apt-get or your package manager.

You can install the other requirements with:

    pip install -r requirements.txt

Examples

With fruits.txt :

You can generate a sankey's diagram with this code (colorDict is optional):

import pandas as pd
import matplotlib.pyplot as plt

from pysankey import sankey

df = pd.read_csv(
    'pysankey/tests/fruits.txt', sep=' ', names=['true', 'predicted']
)
colorDict = {
    'apple':'#f71b1b',
    'blueberry':'#1b7ef7',
    'banana':'#f3f71b',
    'lime':'#12e23f',
    'orange':'#f78c1b',
    'kiwi':'#9BD937'
}

ax = sankey(
    df['true'], df['predicted'], aspect=20, colorDict=colorDict,
    leftLabels=['banana','orange','blueberry','apple','lime'],
    rightLabels=['orange','banana','blueberry','apple','lime','kiwi'],
    fontsize=12
)

plt.show() # to display
plt.savefig('fruit.png', bbox_inches='tight') # to save

With customer-goods.csv :

,customer,good,revenue
0,John,fruit,5.5
1,Mike,meat,11.0
2,Betty,drinks,7.0
3,Ben,fruit,4.0
4,Betty,bread,2.0
5,John,bread,2.5
6,John,drinks,8.0
7,Ben,bread,2.0
8,Mike,bread,3.5
9,John,meat,13.0

You could also weight:

import pandas as pd
import matplotlib.pyplot as plt

from pysankey import sankey

df = pd.read_csv(
    'pysankey/tests/customers-goods.csv', sep=',',
    names=['id', 'customer', 'good', 'revenue']
)
weight = df['revenue'].values[1:].astype(float)

ax = sankey(
      left=df['customer'].values[1:], right=df['good'].values[1:],
      rightWeight=weight, leftWeight=weight, aspect=20, fontsize=20
)

plt.show() # to display
plt.savefig('customers-goods.png', bbox_inches='tight') # to save

Similar to seaborn, you can pass a matplotlib Axes to sankey function with the keyword ax=:

import pandas as pd
import matplotlib.pyplot as plt

from pysankey import sankey

df = pd.read_csv(
        'pysankey/tests/fruits.txt',
        sep=' ', names=['true', 'predicted']
)
colorDict = {
    'apple': '#f71b1b',
    'blueberry': '#1b7ef7',
    'banana': '#f3f71b',
    'lime': '#12e23f',
    'orange': '#f78c1b'
}

ax1 = plt.axes()

ax1 = sankey(
      df['true'], df['predicted'], aspect=20, colorDict=colorDict,
      fontsize=12, ax=ax1
)

plt.show()

Important informations

Use of figureName, closePlot and figSize in sankey() has been removed. This is done so matplotlib is used more transparently as this [issue] suggested (https://github.com/anazalea/pySankey/issues/26#issue-429312025) on the original github repo.

Now, sankey() does less of the customization and let the user do it to their liking by returning a matplotlib Axes object, which mean the user also has access to the Figure to customise. Then they can choose what to do with it - showing it, saving it with much more flexibility.

Recommended changes to your code from pySankey

• To save a figure, after sankey(), one can simply do:

  plt.savefig("<figureName>.png", bbox_inches="tight", dpi=150)

• To display the diagram, simply do plt.show() after sankey().

• You can modify the sankey size by changing the one from the matplotlib figure.

    plt.gcf().set_size_inches(figSize)
  

• It is possible to modify the diagram font looks, for example, add the following lines before calling sankey() :

    plt.rc("text", usetex=False)
    plt.rc("font", family="serif")
  

Package development

Lint

pylint pysankey

Testing

python -m unittest

Coverage

coverage run -m unittest
coverage html
# Open htmlcov/index.html in a navigator