Naming Prosecution Mitigator
Miss any of our Open RFC calls?Watch the recordings here! »

stats-accumulator

1.1.2 • Public • Published

stats-accumulator

NPM

A simple tool for calculating incremental stats on numeric streams. Forked from stats-incremental for minimal memory and maximum portability.

E.g. given a source of numbers of unknown length that you would like to at any given time know any of:

  • count
  • min
  • max
  • sum
  • variance
  • stddev

This module can be used either with Node streams via a wrapper such as through2 or without being streaming.

Example

Non-streaming:

var Stats = require("stats-accumulator");
 
var dice = require("dice");
var s = Stats();
 
var rolls = [];
for (var i = 0; i < 100; i++) {
  s.update(dice.sum(dice.roll("2d6")));
  console.log(s.toJSON());
}
 
/* E.g.
  { n: 97,
  min: 2,
  max: 12,
  sum: 673,
  mean: 6.938144329896907,
  variance: 5.851843979168881,
  stddev: 2.419058490233107 }
*/
 
console.log(s.mean);
console.log(s.stddev);

With streams:

var spigot = require("stream-spigot");
var through2 = require("through2");
var terminus = require("terminus");
 
var Stats = require("stats-accumulator");
var s = Stats();
 
var statStream = through2.obj(function (chunk, encoding, callback) {
  s.update(chunk);
  if (s.n % 100000 === 0) {
    console.log(s.toJSON());
  }
  this.push(chunk);
  callback();
});
 
spigot
  .sync({ objectMode: true }, Math.random)
  .pipe(statStream)
  .pipe(terminus.devnull({ objectMode: true }));
 
/*
  { n: 100000,
    min: 2.0884908735752106e-7,
    max: 0.9999937505926937,
    sum: 49861.06196602131,
    mean: 0.49861061966021336,
    variance: 0.08331362954827709,
    stddev: 0.28864100462040576 }
  { n: 200000,
    min: 2.0884908735752106e-7,
    max: 0.9999937505926937,
    sum: 99904.73041411326,
    mean: 0.49952365207056687,
    variance: 0.08316120223669865,
    stddev: 0.2883768406732736 }
*/

API

const Stats = require("stats-accumulator")

var stats = new Stats(smaBins)

Create a new incremental stats aggregator. The smaBins argument is optional (default 50) and will choose the size of recent window to retain to calculate the Simple Moving Average on the recent data.

stats.update(value)

Update the aggregator with a value. Converted to a Number via parseFloat. If this results in NaN the update is skipped.

stats.toJSON()

Get a up-to-date clone of all of the stats stored.

E.g.

{ n: 97,
  min: 2,
  max: 12,
  sum: 673,
  mean: 6.938144329896907,
  variance: 5.851843979168881,
  stddev: 2.419058490233107 }

stats.n

The count of observations.

stats.min

The min value observed.

stats.max

The max value observed.

stats.sum

The sum of all values observed.

stats.mean

The arithmetic mean of the observations.

stats.variance

The variance from the mean.

stats.stddev

The standard deviation of the values from the mean.

Alternatives

stats-lite Operates on complete sets of numbers.

stream-statistics Is a similar module dedicated to streams.

stats-incremental Is the original module, but it accumulates a moving average.

LICENSE

MIT

Install

npm i stats-accumulator

DownloadsWeekly Downloads

0

Version

1.1.2

License

MIT

Unpacked Size

6.95 kB

Total Files

4

Last publish

Collaborators

  • avatar