PyCircStat2: Circular statistics with Python

A rework of pycircstat.

Key Features | Installlation | API Reference | Examples ( Books | Topics )

Key Features

• One-Stop Circular Data Analysis Pipeline with Circular Class

  The Circular class simplifies circular data analysis by providing automatic data transformation, descriptive statistics, hypothesis testing, and visualization tools—all in one place.

  from pycircstat2 import Circular
  data = [30, 60, 90, 120, 150]
  circ = Circular(data, unit="degree")
  print(circ.summary())
  circ.plot(config={"rose": True, "mean": True})
  

• Compatibility with Legacy APIs

  APIs for descriptive statistics and hypothesis testing follow the conventions established by the original circstat-matlab and pycircstat, ensuring ease of use for existing users.

• Wide-Ranging Circular Distributions

  The package supports a variety of circular distributions, including but not limited to:

  • Symmetric distributions: Circular Uniform, Cardioid, Cartwright, Wrapped Normal, Wrapped Cauchy, von Mises (and its flat-top extension), and Jones-Pewsey.
  • Asymmetric distributions: Sine-skewed Jones-Pewsey, Asymmetric Extended Jones-Pewsey, Inverse Batschelet.

Also see the full feature checklist here.

Installation

To install the latest tagged version:

pip install pycircstat2

Or to install the development version, clone the repository and install it with pip install -e:

git clone https://github.com/circstat/pycircstat2
pip install -e pycircstat2

API Reference

The API reference is available here.

Example notebooks

In the notebooks below, we reproduce examples and figures from a few textbooks on circular statistics.

Books

• Statistical Analysis of Circular Data (Fisher, 1993)
• Chapter 26 and 27 from Biostatistical Analysis (Zar, 2010).
• Circular Statistics in R (Pewsey, et al., 2014)

And a few more examples on selective topics:

Topics

1. Utils
2. Descriptive Statistics
3. Hypothesis Testing
4. Circular Models
5. Regression
6. Clustering