vega/datalib

Name: datalib

Owner: Vega

Description: JavaScript data utility library.

Created: 2015-04-14 17:45:56.0

Updated: 2018-01-18 07:23:48.0

Pushed: 2018-01-16 22:56:04.0

Homepage: http://vega.github.io/datalib/

Size: 3205

Language: JavaScript

GitHub Committers

User	Most Recent Commit	# Commits

Other Committers

User	Email	Most Recent Commit	# Commits

README

datalib

Datalib is a JavaScript data utility library. It provides facilities for data loading, type inference, common statistics, and string templates. While datalib was created to power Vega and related projects, it is also a standalone library useful for data-driven JavaScript applications on both the client (web browser) and server (e.g., node.js).

For documentation, see the datalib API Reference.

Use

Datalib provides a set of utilities for working with data. These include:

• Loading and parsing data files (JSON, TopoJSON, CSV, TSV).
• Summary statistics (mean, deviation, median, correlation, histograms, etc).
• Group-by aggregation queries, including streaming data support.
• Data-driven string templates with expressive formatting filters.
• Utilities for working with JavaScript functions, objects and arrays.

Datalib can be used both server-side and client-side. For use in node.js, simply npm install datalib or include datalib as a dependency in your package.json file. For use on the client, install datalib via bower install datalib or include datalib.min.js on your web page. The minified JS file is built using browserify (see below for details).

Example

oad datalib.
dl = require('datalib');

oad and parse a CSV file. Datalib does type inference for you.
he result is an array of JavaScript objects with named values.
arsed dates are stored as UNIX timestamp values.
data = dl.csv('http://vega.github.io/datalib/data/stocks.csv');

how summary statistics for each column of the data table.
ole.log(dl.format.summary(data));

ompute mean and standard deviation by ticker symbol.
rollup = dl.groupby('symbol')
ummarize({'price': ['mean', 'stdev']})
xecute(data);
ole.log(dl.format.table(rollup));

ompute correlation measures between price and date.
ole.log(
.cor(data, 'price', 'date'),      // Pearson product-moment correlation
.cor.rank(data, 'price', 'date'), // Spearman rank correlation
.cor.dist(data, 'price', 'date')  // Distance correlation


ompute mutual information distance between years and binned price.
bin_price = dl.$bin(data, 'price'); // returns binned price values
year_date = dl.$year('date');       // returns year from date field
counts = dl.groupby(year_date, bin_price).count().execute(data);
ole.log(dl.mutual.dist(counts, 'bin_price', 'year_date', 'count'));

Build Process

To use datalib in the browser, you need to build the datalib.js and datalib.min.js files. We assume that you have npm installed.

1. Run npm install in the datalib folder to install dependencies.
2. Run npm run build. This will invoke browserify to bundle the source files into datalib.js, and then uglify-js to create the minified datalib.min.js.

Webpack 1

If you are using Webpack 1, you need to enable a JSON-loader. To do so, first npm install --save json-loader, then add the loader to your webpack config:

dule: {
loaders: [{
  test: /\.json$/,
  loader: 'json-loader'
}]