Creating Sankey Diagrams in Python with Matplotlib and Pyecharts

A Sankey diagram (also called a Sankey energy flow diagram) is a specialized flow chart where the width of each branch corresponds to the magnitude of the flow, making it useful for visualizing energy, material composition, or financial data.

The diagram mainly appears in two styles, which are illustrated below.

Below is a Matplotlib example that builds a simple Sankey chart:

import matplotlib.pyplot as plt
import matplotlib as mpl
import numpy as np
from matplotlib.sankey import Sankey

plt.rcParams['figure.figsize'] = (7.0, 5.0)
plt.rcParams['savefig.dpi'] = 100  # image resolution
plt.rcParams['figure.dpi'] = 100

flows = [0.2, 0.1, 0.4, 0.3, -0.6, -0.05, -0.15, -0.2]
labels = ['', '', '', '', 'family', 'trip', 'education', 'sport']
orientations = [1, 1, 0, -1, 1, -1, 1, 0]

sankey = Sankey()
sankey.add(flows=flows,
           labels=labels,
           orientations=orientations,
           color='c',
           fc='lightgreen',
           patchlabel='Life Cost',
           alpha=0.7)

diagrams = sankey.finish()
diagrams[0].texts[4].set_color('r')
diagrams[0].texts[4].set_weight('bold')
diagrams[0].texts[4].set_fontsize(15)
diagrams[0].text.set_fontsize(20)
diagrams[0].text.set_fontweight('bold')

plt.title('The Flow Chart of Daily Cost')
plt.show()

The resulting chart looks like this:

A more visually striking Sankey diagram, similar to the one shown in an ECharts tutorial, is also presented:

To create Sankey diagrams with Python, you can also use the Pyecharts library. The following code demonstrates a vertical Sankey chart with custom colors and interactive tooltips:

from pyecharts import options as opts
from pyecharts.charts import Sankey

colors = ["#67001f","#b2182b","#d6604d","#f4a582","#fddbc7",
          "#d1e5f0","#92c5de","#4393c3","#2166ac","#053061"]

nodes = [
    {"name": "a"},
    {"name": "b"},
    {"name": "a1"},
    {"name": "b1"},
    {"name": "c"},
    {"name": "e"},
]

links = [
    {"source": "a", "target": "a1", "value": 5},
    {"source": "e", "target": "b", "value": 3},
    {"source": "a", "target": "b1", "value": 3},
    {"source": "b1", "target": "a1", "value": 1},
    {"source": "b1", "target": "c", "value": 2},
    {"source": "b", "target": "c", "value": 1},
]

c = (
    Sankey()
    .set_colors(colors)
    .add(
        "sankey",
        nodes=nodes,
        links=links,
        pos_bottom="10%",
        focus_node_adjacency="allEdges",
        orient="vertical",
        linestyle_opt=opts.LineStyleOpts(opacity=0.2, curve=0.5, color="source"),
        label_opts=opts.LabelOpts(position="top"),
    )
    .set_global_opts(
        title_opts=opts.TitleOpts(title="Sankey-Vertical"),
        tooltip_opts=opts.TooltipOpts(trigger="item", trigger_on="mousemove"),
    )
    .render("sankey_vertical.html")
)

The chart produced by this script appears as follows:

When constructing Sankey diagrams, ensure that your data is organized in a source → target → value format so that the relationships between adjacent categories can be correctly visualized.

Disclaimer: This article is compiled from online sources; all copyrights belong to the original authors. If any information is incorrect or infringes rights, please contact us for removal or authorization.