When customers branch into different actions, a Sankey diagram makes the flow easy to follow. For interactive exploration, Plotly’s go.Sankey works well. You can inspect flow volume on hover and share as HTML.
import plotly.graph_objects as go
labels = [
"Entry",
"Free signup",
"Cart abandon",
"FAQ visit",
"Bounce",
"Paid conversion",
"Return visit",
"Churn",
]
sources = [0, 0, 0, 0, 1, 1, 1]
targets = [1, 2, 3, 4, 5, 6, 7]
values = [420, 280, 200, 100, 260, 100, 60]
fig = go.Figure(
data=[
go.Sankey(
arrangement="snap",
node=dict(
pad=18,
thickness=18,
line=dict(color="rgba(0,0,0,0.2)", width=1),
label=labels,
color=[
"#4C78A8",
"#F58518",
"#E45756",
"#72B7B2",
"#54A24B",
"#EECA3B",
"#B279A2",
"#FF9DA6",
],
),
link=dict(
source=sources,
target=targets,
value=values,
color=[
"rgba(76,120,168,0.45)",
"rgba(229,87,86,0.45)",
"rgba(114,183,178,0.45)",
"rgba(84,162,75,0.45)",
"rgba(245,133,24,0.45)",
"rgba(238,202,59,0.45)",
"rgba(178,121,162,0.45)",
],
),
)
]
)
fig.update_layout(
title=dict(text="Customer flow after visit (Plotly Sankey)", x=0.5, font=dict(size=18)),
font=dict(size=12, color="#222"),
margin=dict(l=10, r=10, t=40, b=10),
)
# Interactive HTML for sharing in a browser
fig.write_html("static/images/visualize/advanced/sankey-diagram.html", include_plotlyjs="cdn")
# Static image for docs (requires kaleido: pip install kaleido)
fig.write_image("static/images/visualize/advanced/sankey-diagram.png", scale=2)
plt.show()
Reading tips #
- Thicker bands represent larger flow, so you can spot major drop-offs quickly.
- Follow connected branches to trace bottlenecks to a specific outcome (e.g., paid conversion).
- If branching is too dense, focus on key paths to keep it readable.