statistics

Standard library module View source

Descriptive statistics and distributions for native numeric data.

The API follows CPython 3.14, including weighted means, ranked correlation, regression, and kernel density estimation. Numeric aggregates return float; median_low, median_high, and modes preserve the input element type. Compensated floating-point arithmetic and widened integer intermediates avoid common cancellation and overflow errors, but do not provide Decimal/Fraction arithmetic. Random sampling accepts integer seeds. Remove NaNs before computing order statistics; native floats do not retain Python object identity for NaNs. Finite product overflow uses a fixed-point recovery accumulator. Supplied-center dispersion is scaled, and integer relational statistics use shifted origins; these can retain finite results and precision lost by CPython's float paths.

class StatisticsError

class StatisticsError

View complete class API →

Constructors 1

__init__

__init__(self, message: str = "")

function median

median(data)

function median_low

median_low(data)

function median_high

median_high(data)

function median_grouped

median_grouped(data, interval = 1.0)

function mode

mode(data)

function multimode

multimode(data)

function quantiles

quantiles(data, n: int = 4, method: str = "exclusive")

function mean

mean(data)

function fmean

fmean(data, weights = None)

function geometric_mean

geometric_mean(data)

function harmonic_mean

harmonic_mean(data, weights = None)

function pvariance

pvariance(data, mu = None)

function pstdev

pstdev(data, mu = None)

function variance

variance(data, xbar = None)

function stdev

stdev(data, xbar = None)

function covariance

covariance(xdata, ydata)

function correlation

correlation(xdata, ydata, method: str = "linear")

class LinearRegression

@tuple

class LinearRegression

View complete class API →

Fields 2

slope float

intercept float

function linear_regression

linear_regression(xdata, ydata, proportional: bool = False)

function kde

kde(data, h, kernel: str = "normal", cumulative: bool = False)

Return a kernel density or cumulative distribution estimator.

function kde_random

kde_random(data, h, kernel: str = "normal", seed: Optional[int] = None)

Return a sampler with a private random generator and optional integer seed.

class NormalDist

class NormalDist

View complete class API →

Constructors 1

__init__

__init__(self, mu = 0.0, sigma = 1.0)

Properties 5

mean

@property

mean(self)

median

@property

median(self)

mode

@property

mode(self)

stdev

@property

stdev(self)

variance

@property

variance(self)

Methods 8

pdf

pdf(self, x)

cdf

cdf(self, x)

zscore

zscore(self, x)

inv_cdf

inv_cdf(self, p: float)

quantiles

quantiles(self, n: int = 4)

overlap

overlap(self, other: NormalDist)

samples

samples(self, n: int, seed: Optional[int] = None)

from_samples

from_samples(data)

Special methods 13

__add__

__add__(x1: NormalDist, x2: NormalDist)

__add__(x1: NormalDist, x2: float)

__sub__

__sub__(x1: NormalDist, x2: NormalDist)

__sub__(x1: NormalDist, x2: float)

__mul__

__mul__(x1: NormalDist, x2: float)

__truediv__

__truediv__(x1: NormalDist, x2: float)

__pos__

__pos__(x1: NormalDist)

__neg__

__neg__(x1: NormalDist)

__radd__

__radd__(x1: NormalDist, x2: float)

__rsub__

__rsub__(x1: NormalDist, x2: float)

__rmul__

__rmul__(x1: NormalDist, x2: float)

__eq__

__eq__(x1: NormalDist, x2: NormalDist)

__ne__

__ne__(self, other: NormalDist)

__hash__

__hash__(self)

__repr__

__repr__(self)