visualization server
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
avro-loader/src
css
dashboard/demos
data-server/src
doc/src/jekyll
http-common/src
js
jsdoc
kafka
loader/src/main/scala/nest/sparkle/loader
log4j
logback
project
protocol-test-kit/src/main/scala/nest/sparkle/time/protocol
protocol/src/main
sbt
servers
spark-avro/src/main/scala/nest/sparkle/loader/spark/avro
spark/src
sparkle-store-tests/src
sparkle-tests
sparkle
store-test-kit/src/main
store/src/main
test-kit
util-kafka/src
util-tests/src/test
util/src
.gitignore
LICENSE
README.md
ToDo
getting-started.md
scalastyle-config.xml
version.sbt

README.md

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 and a scala/jvm server library. The client and server are connected by websockets protocol for requesting and delivering 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. See Visualize the Things for a brief introduction to sparkle, lessons on garbage collection with reative streams, and comments on the future of the Lambda Architecture. See The Live Layer for a discussion of the Live Layer, a key component missing from the Lambda Architecture. See Sparkle Google Docs Folder for other docs.

Components

Sparkle contains three core components:

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

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.