statistics

The statistics library provides functions for calculating mathematical statistics of numeric data, following Python’s statistics module: averages, central tendency, and measures of spread. Every function takes a single list of numbers (mode also accepts other values).

Available Functions

Example

import statistics

data = [2, 4, 4, 4, 5, 5, 7, 9]

print(statistics.mean(data), statistics.fmean([1, 2, 3]))  # 5.0 2.0
print(statistics.median(data), statistics.median([1, 3, 5]))  # 4.5 3
print(statistics.mode(["a", "b", "b"]))           # b
print(round(statistics.geometric_mean([1, 2, 4, 8]), 3))  # 2.828
print(round(statistics.harmonic_mean([40, 60]), 6))  # 48.0

# Sample statistics (data is a sample of a larger population)
print(round(statistics.variance(data), 3), round(statistics.stdev(data), 3))  # 4.571 2.138
# Population statistics (data is the whole population)
print(statistics.pvariance(data), statistics.pstdev(data))  # 4.0 2.0

Differences from Python

• data must be a list; tuples, ranges and other iterables raise a type error, so wrap them with list(...).
• mean() and pvariance() always return a float (Python returns an int for integer data that divides evenly, e.g. mean([1, 2, 3]) is 2.0 here).
• variance()/stdev() take no xbar argument, and pvariance()/pstdev() take no mu argument.
• geometric_mean() and harmonic_mean() raise on zero or negative values (Python returns 0 for a zero).
• median_low, median_high, median_grouped, multimode, quantiles, and NormalDist are not available.

See Also

• math: mathematical functions and constants.
• random: random number generation, including Gaussian and other distributions.