Freedom251 › Statistical Analysis Software
PyUnfold
What PyUnfold is
PyUnfold is a free, MIT-licensed Python package for iterative unfolding, a deconvolution method that estimates true distributions from measured data. Its Bayesian approach provides uncertainty estimates for statistical and systematic errors. The package supports multidimensional unfolding, custom initial priors, optional smoothing with tunable univariate splines, and an adjustable test-statistic criterion for convergence. It is designed for scientific users, including first-time users and experimental analyses. PyUnfold supports statistical tests, Bayesian analysis, and scripting, but not regression models. Installation is available through pip or conda, with NumPy, SciPy, pandas, and six listed as dependencies. Documentation includes notebooks covering the basic API, custom priors, spline regularization, and multivariate unfolding, launchable with Binder. It runs on Linux desktop systems. PyPI lists version 0.5.0, released November 19, 2018.
Who it suits
PyUnfold suits scientific programmers and experimental analysts who need Bayesian unfolding, uncertainty estimates, custom priors, or multidimensional analysis in Python.
What it does well
- Free MIT-licensed Python package
- Estimates statistical and systematic uncertainties
- Supports multidimensional unfolding
- Allows custom priors and spline regularization
- Installable with pip or conda
What to know first
- Does not provide regression models
- Linux desktop platform listed
- Latest listed release is from 2018
Plans
| Plan | Billed and limits | Price |
|---|---|---|
| PyUnfold | Open-source Python package · MIT license | Free |
As jrbourbeau.github.io lists them, read 11 Oct 2026. The maker's own page has the final word.
Verdict
PyUnfold offers configurable Bayesian unfolding with uncertainty handling and scientific-computing integrations. Its older listed release and lack of regression models are important considerations for new projects.
Facts from PyUnfold's own pages
- Bayesian analysis
- Yes · jrbourbeau.github.io, 23 Sept 2026
- Convergence
- Users can adjust the test statistic criterion used to compare successive unfolded distributions and determine convergence. · jrbourbeau.github.io, 11 Oct 2026
- Custom priors
- Users can supply custom initial prior distributions. · jrbourbeau.github.io, 11 Oct 2026
- Deployment
- desktop · jrbourbeau.github.io, 23 Sept 2026
- Examples
- The documentation includes notebooks for a basic API tutorial, custom priors, spline regularization, and multivariate unfolding, which can be launched with Binder. · jrbourbeau.github.io, 11 Oct 2026
- Free plan
- Yes · jrbourbeau.github.io, 23 Sept 2026
- Installation requirement
- The installation page lists Python >= 2.7 or >= 3.4 as a dependency requirement. · jrbourbeau.github.io, 11 Oct 2026
- Integrations
- The documented scientific computing dependencies are NumPy, SciPy, pandas, and six; the package can be installed with pip or conda. · jrbourbeau.github.io, 11 Oct 2026
- Intended users
- The package is designed for members of scientific disciplines and for first-time users as well as experimental analyses. · jrbourbeau.github.io, 11 Oct 2026
- License
- The public GitHub repository identifies PyUnfold as MIT licensed. · github.com, 11 Oct 2026
- Method
- Its iterative unfolding method uses Bayes’ theorem and provides uncertainty estimates that account for statistical and systematic errors. · jrbourbeau.github.io, 11 Oct 2026
- Multidimensional analysis
- PyUnfold supports multidimensional unfolding. · jrbourbeau.github.io, 11 Oct 2026
- Purpose
- PyUnfold is a Python package for iterative unfolding, a deconvolution method used to estimate true distributions from measured data. · jrbourbeau.github.io, 11 Oct 2026
- Regression models
- No · jrbourbeau.github.io, 23 Sept 2026
- Regularization
- Optional regularization uses tunable univariate splines to smooth unfolded distributions. · jrbourbeau.github.io, 11 Oct 2026
- Release
- PyPI lists version 0.5.0, released November 19, 2018, as the latest version. · pypi.org, 11 Oct 2026
- Research use
- The project says PyUnfold has been used for cosmic-ray energy spectrum measurement at HAWC and cosmic-ray composition analysis using IceCube. · jrbourbeau.github.io, 11 Oct 2026
- Scripting support
- Yes · jrbourbeau.github.io, 23 Sept 2026
- Statistical tests
- Yes · jrbourbeau.github.io, 23 Sept 2026
- Support
- The project directs questions, feature proposals, and bug reports to its GitHub issue tracker. · jrbourbeau.github.io, 11 Oct 2026
Where PyUnfold stands on Freedom251
- Statistical Analysis Software#62 of 106