Mint Statistics

A lightweight and fast statistics library for JavaScript and TypeScript.

Why Mint Stats?

• 5-10x faster than traditional implementations (Quickselect for median)
• Fully typed - Built with TypeScript from the ground up
• Zero dependencies - Size: 3.7KB
• Tree-shakeable - Import only what you need

Install

npm install mintstats

Quick Start

import statistics from "mintstats";

const scores = [85, 92, 78, 90, 88];

statistics.mean(scores);    // 86.6
statistics.median(scores);  // 88
statistics.stdev(scores);   // 5.22

// Works with objects too
const students = [
  { name: "Alice", score: 85 },
  { name: "Bob", score: 92 },
];

statistics.mean(students, "score");             // 88.5
statistics.percentile(students, 75, "score");   // 90.5

API

Basic Stats

• mean(arr, key?) - Average
• median(arr, key?) - Middle value (uses Quickselect - O(n))
• mode(arr, key?) - Most frequent value(s)
• range(arr, key?) - Max - Min

Variability

• variance(arr, options?) - Variance (sample or population)
• stdev(arr, options?) - Standard deviation
• percentile(arr, p, key?) - Nth percentile (0-100)

Options for variance/stdev:

{
  key: 'propertyName',  // for object arrays
  isSample: true        // sample (n-1) or population (n)
}

Performance

Benchmark on 10,000 items:

Traditional (sort): 588 ops/sec
Mint Stats:        3,438 ops/sec  -> 5.85x faster

TypeScript

Full type safety with automatic inference:

statistics.mean([1, 2, 3]);         // Works
statistics.mean(objects, "value");  // Works
statistics.mean(objects);           // Type error - key required

License

MIT © Peakk