�ݺ�ߣ

Copyright ? 2014 Intridea Inc. All rights reserved.
Visualizing Big Data on Small Devices
Tom Zeng
Director of Engineering
tom@intridea.com
@tomzeng
www.linkedin.com/in/tomzeng

Agenda
Introduction
Front End - HTML5/Bootstrap, Backbone/Co?eeScript, D3, MapBox
Backend - Rails, MongoDB
Big Data Processing - Hadoop, Hive, Pig
Showcase - Mobile and Data Visualization Related Projects
Q & A

Intridea - Rails, UX/Data Visualization, Mobile, Big Data, e-commerce
American Bible Society (ABS http://www.americanbible.org/) - partners with Bible publishers
�� Provides API access to 539 Bible versions in 242 languages
�� The usage of the APIs is tracked at the verse level, along with ip location, timestamp, and duration
�� 530 million view logs/year(��12-��13 data), 1.5 mil/day, each view log has packed about 12 bible views
�� Amounts to 5-6 billion Bible views each year
ABS asked Intridea to build the dashboard app Scripture Analytics (http://www.scriptureanalytics.com)
Introduction

BY WORKING REMOTELY
9,816 Hours Saved Annually
ACROSS THE US & OVERSEAS
30+ Employees
FOUNDED & STARTED IN 2007
Washington D.C.
We Make
! ? #
ON GITHUB
Open Source Software
$

Major Open Source Contributions
OmniAuth is a flexible authentication
system utilizing Rack middleware.
OmniAuth
An opinionated micro-framework for
creating REST-like APIs in Ruby.
Grape
Hashie is a simple collection of useful
Hash extensions.
Hashie
A Ruby wrapper for the OAuth 2.0
protocol.
oauth2
A symbol font that makes it easy to
create a map of the U.S. with HTML/CSS.
Stately
A generic swappable back-end for JSON
handling.
Multi_JSON
Mission control dashboard for your
distributed teams.
Houston
github.com/intridea

INTRIDEA
Simplified and modernized application
experience for ADP
Application engineering for
BusinessWeek.com
World��s first 100% web based Point of
Sale system

Where people go to make a di?erence
with their investment capital.
Simplifying ADP��s core business:
Payroll
Engineering for the most-tra?icked
wedding planning solution

ABS Scripture Analytics Query Requirements
Visualizations

Public and private dashboards
visualizing Bible reading
across the Internet
Up to the minute dashboards showing what Bible
verses are being read when, and where, all over
the globe.

Mobile
Desktop
Tablet
Responsive Web App
www.scriptureanalytics.com

Front End User Interface
Single Page Application using Backbone.js
Co?eeScript (Ruby like, Jasmine in Co?eeScript similar to RSpec)?
D3 for Data Visualization
Twitter Bootstrap for Responsive UI
Packery for Responsive Layout - http://packery.metafizzy.co/
Mapbox for Map Rendering - https://www.mapbox.com/

Backend Servers/Services
Rails on Ruby Application mostly as the API server
MongoDB as the data store/cache
Mongoid for Active Record like queries
MongoDB Aggregation Framework for complex queries
Pulling data periodically from S3 to populate the Mongo database
Local R&D Hadoop and Mongo clusters for data exploration

MongoDB
Document oriented, schema free, JSON format
Very high data read and write throughput
Rich query capabilities (aggregation framework), flexible indexes
Scale with auto-sharded replica sets
Map/Reduce in JavaScript

Hadoop/Pig/Hive/Impala
Hadoop cluster (AWS Elastic Map/Reduce on-demand) to process and store data in S3
Pig to parse, transform, geo-code data
Hive to query data and generate aggregated JSON reports
Impala is similar to Hive (but much fast than the older version of Hive), used for ETL
!

Elastic Map/Reduce Hadoop Cluster - On Demand Processing

Elastic Map/Reduce Hadoop Cluster - Terminated when done

Cloudera CDH4 - on local 10-node cluster

Cloudera CDH4 - Streaming Data into Hive Table

Pig Sample Query

Hive Sample Query

Hive Query Results - Bible views by City

Hive Query Results - Most popular verse before Mother��s day
28
Her children show their appreciation, and her husband praises her.

Hive Query Results - Most popular verse on Mother��s day
28

Hive Query Results - Most popular verse after Mother��s day
28

MongoDB Aggregation Framework Example
http://docs.mongodb.org/manual/core/aggregation-pipeline/

MongoDB Aggregation Framework Example

ABS Data Processing using Hadoop and MongoDB

Analyzing Twitter using Hadoop and MongoDB

Mobile and Data Visualization Project Showcase
ADP
BLiNQ
PEW Templeton - Global Religious Futures
Cato Institute - HumanProgress
!
!

Redefining ADP��s
touch and desktop
experiences
ADP processes one out of every six
paychecks in the United States. We��re
bringing payroll into the decade of touch.

ADP TLM

ADP HCR

INTRIDEA

Dashboard insights for 600 of
the world��s largest advertisers
Analytics, planning, and flight management for social
advertising campaigns and brand engagement.

INTRIDEA
INTRIDEA

Analyzing religious
change and its impact
on societies around the
world
Interactive website exploring the patterns
and trends in religions across the globe

PEW Global Research
PEW Global Research

Human advancement to
a higher stage
Human Progress seeks to document changes in
living standards in the past and present while
explaining and exploring the best ways to improve
conditions for people.

INTRIDEA
humanprogress.org

Gracias
Merci
��꤬�Ȥ�
Danke
лл
Thank You
Tom Zeng
Director of Engineering
tom@intridea.com
@tomzeng
www.linkedin.com/in/tomzeng

�ݺ�ߣ

RubyNation Visualizing Big Data on Small Devices

More Related Content

RubyNation Visualizing Big Data on Small Devices