201503 ieg preso_ramesh bhashyam_teradata_publish

Information Excellence 2015 Spring Summit Business Analytics

Ramesh Bhashyam, CTO. Teradata R&D LabsUniversal Big Data Architecture

Information Excellence 2015 Spring Summit Business Analytics

Big Data Architecture Information Excellence Summit

Ramesh Bhashyam Teradata R&D Labs

[email protected] 28 Feb 2015

2

The business value of data and analytics are no longer in question…

Copyright Teradata Corporation 2011

We are used to the idea of deploying new technology to

improve productivity and efficiency... But data are no

longer merely the by-product of process improvement, they are becoming the raw material of

business.

3

• More data has been created in the last three years than in all past 40,000 years.

• Total data has quadrupled in the last three years.

• Driven by several trends: Moore’s law, web, social media, sensors, smart phones, cameras, network of things.

• Unit of Analysis has changed

Increasing Complexity of Data

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

Volu

me

in E

xaby

te

Transactional Data

Documents Web Data Sensor Data

1990 2000 2010 20202014

Transactions

Interactions

Observations

4 © 2014 Teradata

Data Plateau

Source: Hortonworks Corporation

5

..and the “Internet of Things” is upon us

“…the main goal of smart systems is to close the loop…

this means using the knowledge gleaned from data to optimize and automate all kinds of processes… ranging

from manufacturing to heading-off car collisions.”

6

Complex Dimensions - Often not Considered

• Value density

– Measure of the extent to which data has to be processed based on form

• Analytic Agility

– Analytics on structured for BI versus unstructured

• Concurrency of Access

– Frequently used data versus infrequently used data or one time data

7

• BI. ROI, reporting and dashboards

• Descriptive analytics. Dividing customers into segments and profiling profitable customers

• Predictive analytics. Future profitability of a customer, scoring models, and predicting outcomes such as churn

• Optimization. Optimization problems and what-if scenarios

=> Acquire, Prepare, Analyze, Validate, and operationalize

Data Driven Decision Making

8

Need to Store All Data

Courtesy: http://med.cornell.libguides.com/HINF5008

9

• Data Lake or Data Platform

• Data R&D

• Data Product

• All Data– Combine behavior with transactions and other customer data

Deriving Value From Data

10

Mind The Gap

• Easy

• Iterative & Fast

• Integrates well with BI/Viz Tools

SQL

• Powerful

• Batch-oriented

• Requires lots of coding

MapReduce

• Huge Opportunity

• Many Players

• Emerging Market

• Hybrid Products

Market WhiteSpace

Math and Stats

DataMining

BusinessIntelligence

Applications

Languages

Marketing

ANALYTIC TOOLS & APPS

USERS

DISCOVERY PLATFORM

DATA WAREHOUSE

ERP

SCM

CRM

Images

Audio and Video

Machine Logs

Text

Web and Social

SOURCES

DATAPLATFORM

ACCESSMANAGEMOVE

UNIFIED DATA ARCHITECTURE

MarketingExecutives

OperationalSystems

FrontlineWorkers

CustomersPartners

Engineers

DataScientists

BusinessAnalysts

Fast Loading

Persistent Refinement

Data Lake/Data Hub

ExploratoryAnalytics

Business Intelligence

Predictive Analytics

Operational Intelligence

Data Discovery

Path, graph, time-series analysis

Pattern Detection

Technology Independent

12

Putting it all Together: Predictive Part Failure AnalysisDefine event-based

scores patterns

Operational analysis

Create Predictive

Models

CAPTURE | STORE | REFINE

Aircraft Sensor Data

Maintenance Records

Identify event paths that lead to part failure and

safety issues over time.

DiscoveryPlatform EDW

Preventative Maintenance

Predictive Safety Warnings

Predictive part lifespan analysis &

optimization

SQL-H SQL-H

Access Hadoop data using SQL to join the data with

Teradata tables

13

Questions?

Information Excellence 2015 Spring Summit Business Analytics

Community Focused

Volunteer DrivenKnowledge Share

Accelerated Learning

Collective ExcellenceDistilled Knowledge

Shared, Non Conflicting Goals

Validation / Brainstorm platformMentor, Guide, Coach

Satisfied, Empowered Professional

Richer Industry and Academia

About Information Excellence Group

Progress Information Excellence

Towards an Enriched Profession, Business and Society

About Information Excellence GroupReach us at: blog: http://informationexcellence.wordpress.com/

linked in: http://www.linkedin.com/groups/Information-Excellence-3893869

Facebook: http://www.facebook.com/pages/Information-excellence-group/171892096247159

presentations: http://www.slideshare.net/informationexcellence

twitter: #infoexcel

email: [email protected]@gmail.com

Have you enriched yourself by contributing to the community Knowledge Share..