Beautiful UpSet plots. Straight from your DataFrame.
Turn overlapping groups into a picture you can understand, customize, and share.

Get started · Try the notebook · Step-by-step guide · Share an idea

Your groups. Their intersections. The full picture.

Which customers use both web and mobile? Which genes appear across conditions? How do your research cohorts overlap?

PyUpSet puts set sizes, exact intersections, and percentages in one clear chart — using the pandas and matplotlib workflow you already know.

Built for the way you work

• DataFrame in, chart out. Pass your membership columns and start exploring.
• Readable by default. Aligned panels, degree colors, counts, and percentages.
• Make it yours. Standard matplotlib objects, custom colors, and PNG or PDF export.
• Keep your data local. Runs in Python or Jupyter on your own machine.

Get started

Requires Python 3.10 or newer.

pip install "git+https://github.com/Up1ng/pyupset.git@v0.1.0"

import pandas as pd
import matplotlib.pyplot as plt
import pyupset

customers = pd.DataFrame({
    "Web":    [1, 1, 1, 0, 1, 0],
    "Mobile": [1, 0, 1, 1, 0, 1],
    "Email":  [1, 1, 0, 1, 0, 0],
    "Paid":   [1, 0, 1, 0, 0, 1],
})

fig, axes = pyupset.plot(customers, sets=list(customers.columns), sort_by="size")
plt.show()

Use 0/1 or True/False columns: one row per observation, one column per group.

From exploration to your next report

Sort by intersection size with sort_by="size", or by the number of groups with sort_by="degree". Adjust the chart with familiar matplotlib commands:

axes["intersections"].set_xlabel("Customers")
fig.savefig("customer-overlap.png", dpi=300, bbox_inches="tight")
fig.savefig("customer-overlap.pdf", bbox_inches="tight")

Follow the step-by-step guide, explore the full reference, or open the quickstart notebook.

Build with us

PyUpSet is open source and at the beginning of its journey. Have a dataset, a use case, or an idea for a better chart? Open an issue — we'd like to hear it.

Star the repository to follow the project. Contributions are welcome: here's how to get involved.

Built with pandas, NumPy & matplotlib. Released under the MIT license.