brycebaril / node-stats-lite

A light statistical package that operates on Arrays.
MIT License
74 stars 11 forks source link

stats-lite

NPM

A fairly light statistical package. Works with numeric arrays, and will automatically filter out non-numeric values and attempt to convert string numeric values.

Install

 npm install stats-lite --save

Example

Live Demo using Browserify!

var stats = require("stats-lite")

var dice = require("dice")

var rolls = []
for (var i = 0; i < 3000; i++) {
  rolls.push(dice.sum(dice.roll("2d6")))
}

console.log("sum: %s", stats.sum(rolls))
console.log("mean: %s", stats.mean(rolls))
console.log("median: %s", stats.median(rolls))
console.log("mode: %s", stats.mode(rolls))
console.log("variance: %s", stats.variance(rolls))
console.log("standard deviation: %s", stats.stdev(rolls))
console.log("sample standard deviation: %s", stats.sampleStdev(rolls))
console.log("85th percentile: %s", stats.percentile(rolls, 0.85))
console.log("histogram:", stats.histogram(rolls, 10))

/* Your exact numbers may vary, but they should be pretty similar:
sum: 21041
mean: 7.0136666666666665
median: 7
mode: 7
variance: 5.8568132222220415
standard deviation: 2.4200853749861886
sample standard deviation: 2.4204888234135953
85th percentile: 10
histogram { values: [ 94, 163, 212, 357, 925, 406, 330, 264, 164, 85 ],
  bins: 10,
  binWidth: 1.05,
  binLimits: [ 1.75, 12.25 ] }
*/

Compatibility Notice: Version 2.0.0+ of this library use features that require Node.js v4.0.0 and above

API

All of the exported functions take vals which is an array of numeric values. Non-numeric values will be removed, and string numbers will be converted to Numbers.

NOTE: This will impact some operations, e.g. mean([null, 1, 2, 3]) will be calculated as mean([1, 2, 3]), (e.g. 6 / 3 = 2, NOT 6 / 4 = 1.5)

numbers(vals)

Accepts an array of values and returns an array consisting of only numeric values from the source array. Converts what it can and filters out anything else. e.g.

numbers(["cat", 1, "22.9", 9])
// [1, 22.9, 9]

sum(vals)

Sum the values in the array.

mean(vals)

Calculate the mean average value of vals.

median(vals)

Calculate the median average value of vals.

mode(vals)

Calculate the mode average value of vals.

If vals is multi-modal (contains multiple modes), mode(vals) will return a ES6 Set of the modes.

variance(vals)

Calculate the variance from the mean for a population.

stdev(vals)

Calculate the standard deviation of the values from the mean for a population.

sampleVariance(vals)

Calculate the variance from the mean for a sample.

sampleStdev(vals)

Calculate the standard deviation of the values from the mean for a sample.

percentile(vals, ptile)

Calculate the value representing the desired percentile (0 < ptile <= 1). Uses the Estimation method to interpolate non-member percentiles.

histogram(vals[, bins])

Build a histogram representing the distribution of the data in the provided number of bins. If bins is not set, it will choose one based on Math.sqrt(vals.length). Data will look like:

histogram {
  values: [ 86, 159, 253, 335, 907, 405, 339, 270, 146, 100 ],
  bins: 10,
  binWidth: 1.05,
  binLimits: [ 1.75, 12.25 ]
}

Where values are the bins and the counts of the original values falling in that range. The ranges can be calculated from the binWidth and binLimits. For example, the first bin values[0] in this example is from 1.75 < value <= 2.8. The third bin values[2] would be 1.75 + (1.05 * 2) < value <= 1.75 + (1.05 * 3) or 3.85 < value <= 4.9.

LICENSE

MIT