scipy.stats.circstd#

scipy.stats.circstd(samples, high=6.283185307179586, low=0, axis=None, nan_policy='propagate', *, normalize=False, keepdims=False)[source]#

Compute the circular standard deviation for samples assumed to be in the range [low to high].

Parameters:

samplesarray_like

Input array.

highfloat or int, optional

High boundary for the sample range. Default is 2*pi.

lowfloat or int, optional

Low boundary for the sample range. Default is 0.

normalizeboolean, optional

If True, the returned value is equal to sqrt(-2*log(R)) and does not depend on the variable units. If False (default), the returned value is scaled by ((high-low)/(2*pi)).

axisint or None, default: None

If an int, the axis of the input along which to compute the statistic. The statistic of each axis-slice (e.g. row) of the input will appear in a corresponding element of the output. If None, the input will be raveled before computing the statistic.

nan_policy{‘propagate’, ‘omit’, ‘raise’}

Defines how to handle input NaNs.

propagate: if a NaN is present in the axis slice (e.g. row) along which the statistic is computed, the corresponding entry of the output will be NaN.
omit: NaNs will be omitted when performing the calculation. If insufficient data remains in the axis slice along which the statistic is computed, the corresponding entry of the output will be NaN.
raise: if a NaN is present, a ValueError will be raised.

keepdimsbool, default: False

If this is set to True, the axes which are reduced are left in the result as dimensions with size one. With this option, the result will broadcast correctly against the input array.

Returns:

circstdfloat: Circular standard deviation.

See also

circmean: Circular mean.
circvar: Circular variance.

Notes

This uses a definition of circular standard deviation from [1]. Essentially, the calculation is as follows.

import numpy as np
C = np.cos(samples).mean()
S = np.sin(samples).mean()
R = np.sqrt(C**2 + S**2)
l = 2*np.pi / (high-low)
circstd = np.sqrt(-2*np.log(R)) / l

In the limit of small angles, it returns a number close to the ‘linear’ standard deviation.

Beginning in SciPy 1.9, np.matrix inputs (not recommended for new code) are converted to np.ndarray before the calculation is performed. In this case, the output will be a scalar or np.ndarray of appropriate shape rather than a 2D np.matrix. Similarly, while masked elements of masked arrays are ignored, the output will be a scalar or np.ndarray rather than a masked array with mask=False.

References

[1]

Mardia, K. V. (1972). 2. In Statistics of Directional Data (pp. 18-24). Academic Press. DOI:10.1016/C2013-0-07425-7.

Examples

>>> import numpy as np
>>> from scipy.stats import circstd
>>> import matplotlib.pyplot as plt
>>> samples_1 = np.array([0.072, -0.158, 0.077, 0.108, 0.286,
...                       0.133, -0.473, -0.001, -0.348, 0.131])
>>> samples_2 = np.array([0.111, -0.879, 0.078, 0.733, 0.421,
...                       0.104, -0.136, -0.867,  0.012,  0.105])
>>> circstd_1 = circstd(samples_1)
>>> circstd_2 = circstd(samples_2)

Plot the samples.

>>> fig, (left, right) = plt.subplots(ncols=2)
>>> for image in (left, right):
...     image.plot(np.cos(np.linspace(0, 2*np.pi, 500)),
...                np.sin(np.linspace(0, 2*np.pi, 500)),
...                c='k')
...     image.axis('equal')
...     image.axis('off')
>>> left.scatter(np.cos(samples_1), np.sin(samples_1), c='k', s=15)
>>> left.set_title(f"circular std: {np.round(circstd_1, 2)!r}")
>>> right.plot(np.cos(np.linspace(0, 2*np.pi, 500)),
...            np.sin(np.linspace(0, 2*np.pi, 500)),
...            c='k')
>>> right.scatter(np.cos(samples_2), np.sin(samples_2), c='k', s=15)
>>> right.set_title(f"circular std: {np.round(circstd_2, 2)!r}")
>>> plt.show()