---
id: coffea
version: "2026.7.0"
license: BSD-3-Clause
license_treatment: permissive
maintenance: active
---

# coffea — Basic tools and wrappers for enabling not-too-alien syntax when running columnar Collider HEP analysis.

License: permissive · Maintenance: active · Downloads: 657.9K/mo

## What it is and what it does

coffea is a Python framework for columnar analysis of high-energy physics data, designed to make working with ROOT files and complex event structures feel natural to scientists familiar with NumPy and Pandas. It wraps uproot for transparent ROOT file access and awkward-array for handling jagged and nested data structures common in particle physics, allowing analysts to write physics logic in a vectorized, array-oriented style rather than event-by-event loops.

The package also provides histogramming, plotting, and lookup-table utilities needed for physics analysis workflows, plus built-in support for horizontal scaling across diverse compute resources—from a laptop to multi-core servers, clusters, and supercomputers—without requiring changes to analysis code. This is achieved through integration with modern distributed computing frameworks like Dask and Parsl, making it possible to prototype locally and scale to production without rewriting.

Use it for:

- Process and analyze ROOT files from particle detector experiments using columnar operations and NumPy-like syntax.
- Build scalable physics analysis pipelines that run unchanged on a laptop, multi-core server, or computing cluster.
- Apply corrections and transformations to Monte Carlo simulation data to match real experimental measurements.
- Manipulate jagged and nested event structures (e.g., variable-length particle lists per event) efficiently.
- Prototype analysis logic locally and then scale horizontally to process large datasets without code changes.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

coffea provides columnar data analysis tools for high-energy physics (HEP) experiments, enabling efficient manipulation of ROOT file data and event structures using NumPy-like syntax with horizontal scaling support.

Yes, if you are doing high-energy physics analysis with ROOT files and want columnar, vectorized workflows with built-in scaling. The package is actively maintained, has no known vulnerabilities, and carries a permissive license. The 25 dependencies are standard for scientific Python and will not be unusual in a physics environment. The Beta status and 'prototype' framing suggest the API may evolve, but active maintenance and recent releases indicate serious ongoing development.

## Install

pip install coffea
uv add coffea
poetry add coffea

## Installing coffea

Before you install: Low friction installation with a pure-wheel distribution. Active maintenance with a recent release 39 days ago and commits through August 2026. The package carries 25 runtime dependencies including heavy scientific stacks (NumPy, SciPy, Numba, Pandas, PyArrow), which are standard for physics analysis but will add meaningful install time.

License in practice: BSD-3-Clause permissive license allows commercial and private use with attribution. No licensing restrictions on derivative works or distribution.

Quickstart:

pip install coffea

import coffea
from coffea.nanoevents import NanoEventsFactory

events = NanoEventsFactory.from_root("file.root").events()

Requires Python 3.10 or later. ROOT file access requires uproot and awkward-array, which are installed as dependencies.

Verify before relying:

- Specific performance characteristics when scaling across different resource types (laptop to cluster to supercomputer)
- Current stability guarantees given the 'Beta' development status and 'prototype' designation in documentation
- Compatibility with specific ROOT file formats and versions beyond the basic uproot support

## Package facts

- License: BSD-3-Clause (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 657.9K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

HEP physics data analysis, columnar ROOT file processing, particle physics event analysis, distributed physics computation, awkward array physics workflows, collider experiment analysis, scalable physics data pipelines, hep-physics, distributed-computing, data-analysis

[View on SkillFed](https://skillfed.io/packages/coffea) · [View on PyPI](https://pypi.org/project/coffea/)