Metadata-Version: 2.1
Name: runningstatistics
Version: 0.0.1
Summary: A simple module for computing the sample statistics of an array in a running/online fashion.
Author-email: Jonathan Lindbloom <jonathan@lindbloom.com>
Project-URL: Homepage, https://github.com/jlindbloom/running-stats
Project-URL: Bug Tracker, https://github.com/jlindbloom/running-stats/issues
Keywords: statistics
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Topic :: Scientific/Engineering :: Mathematics
Requires-Python: >=3.7
Description-Content-Type: text/markdown
License-File: LICENSE

# running-stats
A simple Python module for computing the sample statistics of an array in a running/online fashion (when samples are not kept).

This is just a Python port of [John Cook]'s C++ code [here](https://www.johndcook.com/blog/skewness_kurtosis/). There is also a Julia implementation by John Myles White [here](https://github.com/johnmyleswhite/StreamStats.jl). The only special feature of this code is that we support computing running statistics of arrays living on a GPU via CuPy.

You may also be interested in the Python module [RunStats](https://grantjenks.com/docs/runstats/).

My interest in computing statistics this way comes from summarizing the samples generated by a MCMC sampler when storing all of the samples is infeasible.
