descriptive-statistics

Summary statistics for plain arrays of numbers. No dependencies, ESM only.

Everything is a plain function that takes a number[] and returns a number (or an object for summary). Variance-based functions default to the sample estimator (Bessel's correction, n - 1); pass false as the last argument to switch to the population estimator.

Install

npm install descriptive-statistics

Usage

import { mean, stddev, median, quantile, correlation } from "descriptive-statistics";

mean([1, 2, 3, 4, 5]);        // 3
stddev([2, 4, 4, 4, 5, 5, 7, 9], false); // 2  (population)
median([1, 2, 3, 4]);         // 2.5
quantile([1, 2, 3, 4, 5, 6, 7, 8, 9, 10], 0.25); // 3.25
correlation([1, 2, 3, 4, 5], [2, 4, 6, 8, 10]);  // 1

API

Single series:

• mean(xs)
• variance(xs, sample = true)
• stddev(xs, sample = true)
• min(xs), max(xs), range(xs)
• median(xs)
• quantile(xs, q) — linear-interpolation (type 7) quantile for q in [0, 1]
• iqr(xs) — interquartile range
• skewness(xs) — bias-adjusted Fisher-Pearson skewness (needs ≥ 3 points)
• kurtosis(xs) — bias-corrected excess kurtosis (needs ≥ 4 points; normal ≈ 0)
• summary(xs) — { n, mean, stddev, min, max, median }

Two series:

• covariance(xs, ys, sample = true)
• correlation(xs, ys) — Pearson r in [-1, 1]

License

BSD-3-Clause