jStat

jStat - JavaScript Statistical Library

jStat provides native javascript implementations of statistical functions. Full details are available in the docs. jStat provides more functions than most libraries, including the weibull, cauchy, poisson, hypergeometric, and beta distributions. For most distributions, jStat provides the pdf, cdf, inverse, mean, mode, variance, and a sample function, allowing for more complex calculations.

Using jStat in a Browser

jStat can be used in the browser. The jStat object will be added to the window. For example:

<script src="components/jstat.js"></script> <!-- include jStat, from the CDN or otherwise -->

<script>
...
var jstat = this.jStat(dataset); // jStat will be added to the window
...
data[i]['cum'] = jstat.normal(jstat.mean(), jstat.stdev()).cdf(data[i].x);
...
</script>

CDN

The library is hosted on jsDelivr using the follwing url:

//cdn.jsdelivr.net/jstat/latest/jstat.min.js

Note that 'latest' can be replaced with any released verion of jStat.

Module Loaders

Currently jStat is exposed as j$ and jStat inside an object, rather than exported directly. This may confuse some module loaders, however should be easily remedied with the correct configuration.

NodeJS & NPM

To install via npm:

npm install --save jStat

When loading under Node be sure to reference the child object.

var jStat = require('jStat').jStat;

RequireJS Shim

For RequireJS not only exports but also init function must be specified.

requirejs.config({
  paths: {
    'jstat': 'path/to/jstat/dist/jstat.min'
  },
  shim: {
    jstat: {
      exports: ['j$', 'jStat'],
      init: function () {
        return {
          j$: j$,
          jStat: jStat
        };
      }
    }
  }
});

Build Prerequisites

In order to build jStat, you need to have GNU make 3.8 or later, Node.js 0.2 or later, and git 1.7 or later. (Earlier versions might work OK, but are not tested.)

Windows users have two options:

1. Install msysgit (Full installer for official Git), GNU make for Windows, and a binary version of Node.js. Make sure all three packages are installed to the same location (by default, this is C:\Program Files\Git).
2. Install Cygwin (make sure you install the git, make, and which packages), then either follow the Node.js build instructions or install the binary version of Node.js.

Mac OS users should install Xcode (comes on your Mac OS install DVD, or downloadable from Apple's Xcode site) and http://mxcl.github.com/homebrew/. Once Homebrew is installed, run brew install git to install git, and brew install node to install Node.js.

Linux/BSD users should use their appropriate package managers to install make, git, and node, or build from source if you swing that way.

Building jStat

First, clone a copy of the jStat git repo by running git clone git://github.com/jstat/jstat.git.

To download all necessary libraries run npm install.

Then, to get a complete, minified version of jStat and all documentation, simply cd to the jstat directory and type make. If you don't have Node installed and/or want to make a basic, uncompressed, unlinted version of jstat, use make jstat instead of make.

The built version of jStat will be put in the dist/ subdirectory.

Generate just the documentation by running make doc. Documentation will be placed in dist/docs by default.

To remove all built files, run make clean.

Running Tests

Execute all tests by running make test.

Or if you wish to run a specific test, cd to test/<subdir> and run node <some_test>-test.js.

Get the Code

Both the minified and unminified source are located in the dist/ directory. For those who don't want to build it themselves.

Contribute

jStat is now going to follow most of the v8 JavaScript guidelines. There will be plenty of source that uses the old style, but we're going to work away from that.

Also, we'll be going through and reimplementing a good portion of the code to run faster. Hopefully it won't take too long to get the project on one basic standard.

When submitting pull requests, no need to check in dist/*.js. They'll be recompiled for distribution anyway.

Join the Community

We always like discussion of how to improve jStat. Join us at our mailing list and let us know what you'd like to see. Also come ask questions in the #jstat channel on irc.freenode.net.