Skip to content

rodneykinney/sparkle

 
 

Repository files navigation

Sparkle

This is the 0.5.x version of Sparkle.

Introduction

Sparkle is a small suite of libraries for making interactive data visualizations. Sparkle includes a javascript client library, a scala/jvm server library and a well defined protocol for requesting and serving data for visualization.

As a standalone tool, sparkle gives users an easy way to make visualize data from a directory of .csv files. The graphs are interactively zoomable in a d3 based web interface. With some easy customization, graphs can be aggregated into dashboards, and customized with simple declarative javascript, and data read in from other sources such as Apache Kafka.

See Sparkle Intro Talk for an introduction presentation.

[coming soon: links to example charts and dashboards]

Components

Sparkle contains three core components:

###Sparkle-Time Server sparkle-time is a mini web server that provides HTTP and websocket apis for data visualization. Sparkle-time collects data from an extensible set of inputs (Apache Kafka and .csv/.tsv data files, hadoop and netcat support coming soon). Sparkle-time stores data in Cassandra. Sparkle-time also hosts an extensible set of data transformations that transform data on demand (downsampling for example).

Sparkle-time can be run as a library inside another server, or it can run standalone if no customization is needed.

###sg.js sg.js is a d3.js based javascript library for displaying interactive charts of potentially large data sets. Charts are customizable with declarative javascript, and extensible with d3. sg.js currently offers support for line charts, scatter plots, bar charts, etc. and is easily extensible. All charts support a drag-based zooming UI.

sg.js also includes a dashboard component for constructing pages that aggregate many charts on the same page. Graphs in the dashboard are resizable. The browser's back button works to undo/redo zoom navigation and chart resizing.

sg.js supports the sparkle-time api, enabling zooming charts of potentially huge server hosted data sets, the option of offloading data transformations to the server, and (soon) server pushed updates to locally displayed charts.

###Sparkle Data API To enable interoperation of multiple clients and servers, the sparkle-time server and sg.js javascript library speak a well-defined data protocol: the Sparkle Data Protocol. The protocol allows clients to request data transformations and data streams from a comformant server.

See Sparkle Data Protocol for details of the protocol. See Sparkle Transforms for the current set of built in transforms.

Plans

Sparkle is currently at version 0.5.x. Sparkle 0.6 will:

  • clean out legacy code
  • more fully support the Sparkle Data Protocol between visualization client and visualizaton server.

Many more future features are listed in the todo list

Using sparkle

See Using-sg

Contributing to Sparkle

See Contributing, Building, and javascript component model.

About

visualization server

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Scala 50.0%
  • JavaScript 35.8%
  • CSS 12.2%
  • Other 2.0%