This is an efficient implementation of Friedman's SuperSmoother [1] algorithm in pure Python. It makes use of NumPy for fast numerical computation.

To install the released version, type

$ pip install supersmoother

This will also install numpy if not already installed.

To install the bleeding-edge source, download the source code from http://github.com/jakevdp/supersmoother and type:

$ pip install .

The only package dependency is numpy; scipy and pytest are also required if you wish to run the test suite.

The package includes several example notebooks showing the code in action. You can see these in the examples/ directory, or view them statically on nbviewer

This code has full unit tests implemented using pytest. To install the latest release and run its tests, use:

$ pip install supersmoother[dev]
$ pytest --pyargs supersmoother

To install from source and run the tests from within the source directory, use:

$ pip install .[dev]
$ pytest supersmoother

The package is tested with Python versions 3.10 through 3.14.

supersmoother was created by Jake VanderPlas

If you use this code in an academic publication, please consider including a citation to our work. Citation information in a variety of formats can be found on zenodo.

[1] Friedman, J. H. (1984) A variable span scatterplot smoother. Laboratory for Computational Statistics, Stanford University Technical Report No. 5. (pdf)