43
Big Data, Big Opportunities From Buzz to Biz Presented on December 6, 2013 Christopher Nguyen, PhD Co-Founder & CEO

Big Data, Big Opportunities

Embed Size (px)

Citation preview

Big Data, Big Opportunities From Buzz to Biz

Presented on December 6, 2013

Christopher Nguyen, PhD Co-Founder & CEO

Step 1 Instrument Calibration

do you get when you cross …

Atlantic Titanic

What

About Halfway!!

do you get when you cross …What

Big Data HadoopIf

are not enough

Whatdo we need?

?

WHAT is Big Data ?

Huge Volume

High Velocity

Great Variety

Big Data Problems

“Standard” Definition

Learn from Data

Predict the Unknowns from the Knowns

automatic customer segmentation

BIG DATA= PROBLEMS

BIG DATA + BIG COMPUTE =BIG OPPORTUNITIES

“Alternative” Definition

WHAT’s the big deal?

I’ve had big data since the 60’s

WHAT’s the big deal?

We’ve had the sun for a few thousand years, right?

Big Data: The question isn’t WHAT. It’s WHY.

Competitive Advantages It Brings Holistic business insights See underlying patterns Predict unknowns from knowns Automate decisions

Technology Cost/Benefit Threshold

Why

WHERE did Big Data Technology come from ?

‘98 ‘06 ‘09 ‘11‘04

A Timeline of Big Data Techology

Google Search

Build biggest Index of the

Internet

Jeff Dean

Doug Cutting

MapReduce Paper

Hadoop

Qi Lu

Eric14

Hadoop

MapR Cloudera

Hortonworks

WHAT SHOULD Big Data Analytics LOOK like ?

HOW are people operationalizing Big Data

0%

50%

100%

EMA Research, Operationalizing Big Data

2012 2013

% Respondents with Big Data Projects Already in Operation

Interactive, Ad Hoc Business Query

Insight Discovery on Aggregated Operational Data

Finance

Engineering

Marketing

Sales

Google BigQuery

Employee Engagement with Operational Data goes through the roof

Mobile Ad Platform

Ad Targeting

CTR Prediction

100+ Million Devices

Customer Service Provider

Product Recommendation

Cross-channel User Experience Optimization

Are there Patterns of Big-Data SUCCESS ?

1Have a Data-driven CULTURE

—Jim Barksdale, former Netscape CEO

If we have data, let’s go with that.

If all we have are opinions, let’s go with mine.

User Survey Opinion “We prefer 30-result pages to 10-result pages”

Empirical Data The extra 500ms causes users to search by 25% less

That’s $12 Billion per year!

Google Search Latency Experiment

—Marissa Mayer, Google then-VP

VS. 2Centralized Data Service Bureau

Distributed Self Service Data Tools

Why? Users didn’t ask enough questions.

!

Why not? Friction too high.

Centralized Data Service Bureau

Didn’t work out

Team self collect, analyze, & learn from own data Lower latency to insight Positive feedback loop to improve tools

Distributed, Self-service Data Tools

Watch WHICH pattern your Chief Data Officer chooses

3Think BIG about Big Data Opportunities

Build Me the Biggest Hangars & the

Longest Runways

I will make sure the planes come—Eric Schmidt, Google then-CEO

WHY Think Big about Big Data ?

Big Data + Machine Learning

Algorithms ModelsData+ =Brain WisdomExperiences+ =

Deep Learning Neuron

(Source: http://capone.mtsu.edu/wlangsto/)

Human

Machine

Deep Learning Neural Networks

(Source: http://www.doc.ic.ac.uk/~nd/surprise_96/journal/vol2/cs11/article2.html)

Human

Machine

Reading Digits in Zip Codes Geoff Hinton, Yann Lecun, et al.

Demo !

http://www.cs.toronto.edu/~hinton/adi/index.htm

Google Brain Project 16,000 Computers to Recognize Cats

Application Unsupervised Image Classification

and 19,999 other concepts Andrew Ng et al.

Source: http://www.nytimes.com/2012/06/26/technology/in-a-big-network-of-computers-evidence-of-machine-learning.html

Words are Vectors Mikolov et al.

Source: http://gigaom.com/2013/08/16/were-on-the-cusp-of-deep-learning-for-the-masses-you-can-thank-google-later/

Portugal -

China +

Bejing =

Lisbon !

!!!

Machine Translation Quoc V. Le et al.

Source: http://arxiv.org/pdf/1309.4168v1.pdf

Computing Paradigm Shift

Abstractions from Human Intelligence

Abstractions by Machine Intelligence

Computing Pattern

Explicit High-Level Instructions

Computing Pattern

Implicit Pattern Learning

WhyBig Data + Big Compute

will lead to super-human

Machine Intelligencein 10+ years

SUMMARYWHAT & WHY of Big Data

EXAMPLES & BEST PRACTICES of Big Data

EXCITING FUTURE IMPLICATIONS of Big Data

Thank you!