SlideShare ist ein Scribd-Unternehmen logo
1 von 13
How Banks Can Ride the Big Data Tsunami
Deepak Ramanathan
Practice Lead, Information Management
SAS Asia Pacific (North)



      Copyright © 2011, SAS Institute Inc. All rights reserved.   1
Copyright © 2011, SAS Institute Inc. All rights reserved.   2
   GOING BIG Key Dimensions




                                Analytics
Data
                                                                     Platforms

                     Copyright © 2011, SAS Institute Inc. All rights reserved.
OUR
PERSPECTIVE
              Big Data is RELATIVE not ABSOLUTE



    Big Data (Noun)

     When volume, velocity and variety of data exceeds an
     organization’s storage or compute capacity for accurate
     and timely decision-making




                       Copyright © 2011, SAS Institute Inc. All rights reserved.
THRIVING IN THE BIG DATA ERA


                   VOLUME
                   VARIETY
DATA SIZE




                   VELOCITY
                   VALUE




                         TODAY                                                              THE FUTURE


                                           Copyright © 2011, SAS Institute Inc. All rights reserved.
ADVANCED ANALYTICS
                                                                                                           TEXT ANALYTICS
                                                                                            Finding treasures in unstructured data
                                                                                                  like social media or survey tools
FORECASTING
                                                                                                        that could uncover insights
Leveraging historical data                                                                              about consumer sentiment
to drive better insight into
decision-making
for the future


                                                     INFORMATION
                                                     MANAGEMENT
                                                                                                              OPTIMIZATION
                                                                                                                 Analyze massive
DATA MINING                                                                                                    amounts of data in
Mine transaction databases                                                                                     order to accurately
for data of spending patterns                                                                               identify areas likely to
that indicate a stolen card..                                                                                   produce the most
                                                                                                                 profitable results
                                               STATISTICS
                                Copyright © 2011, SAS Institute Inc. All rights reserved.                                              6
THE ANALYTICS LIFECYCLE


                                                             IDENTIFY /
                                                            FORMULATE
BUSINESS                          EVALUATE /
                                                             PROBLEM                                       BUSINESS
MANAGER                            MONITOR                                           DATA                  ANALYST
                                   RESULTS                                        PREPARATION
Domain Expert                                                                                              Data Exploration
Makes Decisions                                                                                            Data Visualization
Evaluates Processes and ROI                                                                                Report Creation

                              DEPLOY
                              MODEL                                                             DATA
                                                                                             EXPLORATION



IT SYSTEMS /                                                                                               DATA MINER /
MANAGEMENT                       VALIDATE                                                                  STATISTICIAN
                                  MODEL                                           TRANSFORM
Model Validation                                                                   & SELECT                Exploratory Analysis
Model Deployment                                            BUILD
                                                                                                           Descriptive Segmentation
Model Monitoring                                            MODEL                                          Predictive Modeling
Data Preparation



                                       Copyright © 2011, SAS Institute Inc. All rights reserved.
Trends in Platforms
                                                                                $20
                                                                                $18
      Hadoop
                          COST PER TERABYTE                                     $16                      COST PER GIGABYTE
Microsoft PDW                                                                   $14
                                                                                $12
       Oracle
                                                                                $10
   Greenplum                                                                     $8
     Teradata                                                                    $6
                                                                                 $4
      Vertica                                                                    $2
                                                                                $-
                $-   $20,000 $40,000 $60,000 $80,000 $100,000

                         Today    2009



                              COST OF STORAGE, MEMORY, COMPUTING
                                       In 2000 a GB of Disk $17 today < $0.07
                                       In 2000 a GB of Ram $1800 today < $1
                                  In 2009 a TB of RDBMS was $70K today < $ 20K



                                             Copyright © 2011, SAS Institute Inc. All rights reserved.
•       Leverage in-memory architecture via
                                             a dedicated software and hardware
                                             appliance
     BIG DATA                        •       Drive high-performance capabilities
                                             across the analytical lifecycle
    MEETS BIG
                                     •       Achieve insights at breakthrough
    ANALYTICS                                speed before questions become
PRODUCT HIGHLIGHTS                           obsolete
                                     •       Offer a consistent interface for current
                                             SAS analytic users




                     Copyright © 2011, SAS Institute Inc. All rights reserved.
CREDIT DEFAULT RISK ASSESSMENT

                                                                                       APPLICATION SCORING
                                                                                       BEHAVIORAL SCORING
                                                                                       COLLECTION SCORING




                                                              RISK ASSESSMENT
                                 ANALYTICAL
                                 LIFECYCLE




                           Copyright © 2011, SAS Institute Inc. All rights reserved.
CUSTOMER
             TRADITIONAL ANALYTICS PROCESS
CASE STUDY




                                                                                      167 Hours



        DATA            MODEL                               MODEL
     EXPLORATION     DEVELOPMENT                          DEPLOYMENT




                          Copyright © 2011, SAS Institute Inc. All rights reserved.
CUSTOMER
             HIGH PERFORMANCE ANALYTICS PROCESS                                              167 Hours
CASE STUDY




                                                                  DEVELOPMENT
                                           E X P L O R AT I O N




                                                                                DEPLOYMENT
                                                                     MODEL




                                                                                  MODEL
  Bottom-line Impact:




                                                  D ATA
    Tens of Millions
       of Dollars


                                                                                              84
                                                                                              SECONDS




                          Copyright © 2011, SAS Institute Inc. All rights reserved.
C op yr i g h t © 2 0 1 2 , S A S I n s t i t u t e I n c . A l l r i g h t s r es er v e d .
                                                                                                SAS.com

Weitere ähnliche Inhalte

Was ist angesagt? (6)

Oracle Business Intelligence (OBIEE) from an Essbase Perspective
Oracle Business Intelligence (OBIEE) from an Essbase PerspectiveOracle Business Intelligence (OBIEE) from an Essbase Perspective
Oracle Business Intelligence (OBIEE) from an Essbase Perspective
 
BI outsourcing and emerging trends 2012
BI outsourcing and emerging trends 2012BI outsourcing and emerging trends 2012
BI outsourcing and emerging trends 2012
 
IASA 2010 - Enhancing the Agent Experience
IASA 2010 - Enhancing the Agent ExperienceIASA 2010 - Enhancing the Agent Experience
IASA 2010 - Enhancing the Agent Experience
 
Tackling big data with hadoop and open source integration
Tackling big data with hadoop and open source integrationTackling big data with hadoop and open source integration
Tackling big data with hadoop and open source integration
 
Linalis UK introduction
Linalis UK introductionLinalis UK introduction
Linalis UK introduction
 
SQL-H a new way to enable SQL analytics
SQL-H a new way to enable SQL analyticsSQL-H a new way to enable SQL analytics
SQL-H a new way to enable SQL analytics
 

Ähnlich wie Asian Bankers Association, Manila Conference

Big Data Needs Big Analytics
Big Data Needs Big AnalyticsBig Data Needs Big Analytics
Big Data Needs Big Analytics
Deepak Ramanathan
 
Infomation models for agile bi
Infomation models for agile biInfomation models for agile bi
Infomation models for agile bi
Ehtisham Rao
 
Martin Wildberger Presentation
Martin Wildberger PresentationMartin Wildberger Presentation
Martin Wildberger Presentation
Mauricio Godoy
 
SAS aster data big data dc presentation public
SAS aster data big data dc presentation publicSAS aster data big data dc presentation public
SAS aster data big data dc presentation public
Teradata Aster
 
ScaleBase Webinar 8.16: ScaleUp vs. ScaleOut
ScaleBase Webinar 8.16: ScaleUp vs. ScaleOutScaleBase Webinar 8.16: ScaleUp vs. ScaleOut
ScaleBase Webinar 8.16: ScaleUp vs. ScaleOut
ScaleBase
 

Ähnlich wie Asian Bankers Association, Manila Conference (20)

SAS Forum India - SAS Visual Analytics - 'Visualize This!'
SAS Forum India - SAS Visual Analytics - 'Visualize This!'SAS Forum India - SAS Visual Analytics - 'Visualize This!'
SAS Forum India - SAS Visual Analytics - 'Visualize This!'
 
SAS Big Data Forum - Transforming Big Data into Corporate Gold
SAS Big Data Forum - Transforming Big Data into Corporate GoldSAS Big Data Forum - Transforming Big Data into Corporate Gold
SAS Big Data Forum - Transforming Big Data into Corporate Gold
 
Technology Strategies for Big Data Analytics,
Technology Strategies for Big Data Analytics, Technology Strategies for Big Data Analytics,
Technology Strategies for Big Data Analytics,
 
Big Data Needs Big Analytics
Big Data Needs Big AnalyticsBig Data Needs Big Analytics
Big Data Needs Big Analytics
 
101 ab 1345-1415
101 ab 1345-1415101 ab 1345-1415
101 ab 1345-1415
 
EDF2013: Selected Talk: Bryan Drexler: The 80/20 Rule and Big Data
EDF2013: Selected Talk: Bryan Drexler: The 80/20 Rule and Big Data EDF2013: Selected Talk: Bryan Drexler: The 80/20 Rule and Big Data
EDF2013: Selected Talk: Bryan Drexler: The 80/20 Rule and Big Data
 
Farklı Ortamlarda Büyük Veri Kavramı -Big Data by Sybase
Farklı Ortamlarda Büyük Veri Kavramı -Big Data by Sybase Farklı Ortamlarda Büyük Veri Kavramı -Big Data by Sybase
Farklı Ortamlarda Büyük Veri Kavramı -Big Data by Sybase
 
Teradata Big Data London Seminar
Teradata Big Data London SeminarTeradata Big Data London Seminar
Teradata Big Data London Seminar
 
Continuous Intelligence: Staying Ahead with Streaming Analytics
Continuous Intelligence: Staying Ahead with Streaming AnalyticsContinuous Intelligence: Staying Ahead with Streaming Analytics
Continuous Intelligence: Staying Ahead with Streaming Analytics
 
Unified big data architecture
Unified big data architectureUnified big data architecture
Unified big data architecture
 
Infomation models for agile bi
Infomation models for agile biInfomation models for agile bi
Infomation models for agile bi
 
Predictive analytics
Predictive analytics Predictive analytics
Predictive analytics
 
Martin Wildberger Presentation
Martin Wildberger PresentationMartin Wildberger Presentation
Martin Wildberger Presentation
 
SAS aster data big data dc presentation public
SAS aster data big data dc presentation publicSAS aster data big data dc presentation public
SAS aster data big data dc presentation public
 
SAP HANA and Apache Hadoop for Big Data Management (SF Scalable Systems Meetup)
SAP HANA and Apache Hadoop for Big Data Management (SF Scalable Systems Meetup)SAP HANA and Apache Hadoop for Big Data Management (SF Scalable Systems Meetup)
SAP HANA and Apache Hadoop for Big Data Management (SF Scalable Systems Meetup)
 
ScaleBase Webinar 8.16: ScaleUp vs. ScaleOut
ScaleBase Webinar 8.16: ScaleUp vs. ScaleOutScaleBase Webinar 8.16: ScaleUp vs. ScaleOut
ScaleBase Webinar 8.16: ScaleUp vs. ScaleOut
 
Big Data Needs Big Analytics
Big Data Needs Big AnalyticsBig Data Needs Big Analytics
Big Data Needs Big Analytics
 
SAS Forum India: Delivering forward-looking insights to drive breakthrough bu...
SAS Forum India: Delivering forward-looking insights to drive breakthrough bu...SAS Forum India: Delivering forward-looking insights to drive breakthrough bu...
SAS Forum India: Delivering forward-looking insights to drive breakthrough bu...
 
Predictive Analytics: Advanced techniques in data mining
Predictive Analytics: Advanced techniques in data miningPredictive Analytics: Advanced techniques in data mining
Predictive Analytics: Advanced techniques in data mining
 
Making New Product Launches Successful: ERP and Supply Chain Strategies
Making New Product Launches Successful: ERP and Supply Chain StrategiesMaking New Product Launches Successful: ERP and Supply Chain Strategies
Making New Product Launches Successful: ERP and Supply Chain Strategies
 

Asian Bankers Association, Manila Conference

  • 1. How Banks Can Ride the Big Data Tsunami Deepak Ramanathan Practice Lead, Information Management SAS Asia Pacific (North) Copyright © 2011, SAS Institute Inc. All rights reserved. 1
  • 2. Copyright © 2011, SAS Institute Inc. All rights reserved. 2
  • 3. GOING BIG Key Dimensions Analytics Data Platforms Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 4. OUR PERSPECTIVE Big Data is RELATIVE not ABSOLUTE Big Data (Noun) When volume, velocity and variety of data exceeds an organization’s storage or compute capacity for accurate and timely decision-making Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 5. THRIVING IN THE BIG DATA ERA VOLUME VARIETY DATA SIZE VELOCITY VALUE TODAY THE FUTURE Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 6. ADVANCED ANALYTICS TEXT ANALYTICS Finding treasures in unstructured data like social media or survey tools FORECASTING that could uncover insights Leveraging historical data about consumer sentiment to drive better insight into decision-making for the future INFORMATION MANAGEMENT OPTIMIZATION Analyze massive DATA MINING amounts of data in Mine transaction databases order to accurately for data of spending patterns identify areas likely to that indicate a stolen card.. produce the most profitable results STATISTICS Copyright © 2011, SAS Institute Inc. All rights reserved. 6
  • 7. THE ANALYTICS LIFECYCLE IDENTIFY / FORMULATE BUSINESS EVALUATE / PROBLEM BUSINESS MANAGER MONITOR DATA ANALYST RESULTS PREPARATION Domain Expert Data Exploration Makes Decisions Data Visualization Evaluates Processes and ROI Report Creation DEPLOY MODEL DATA EXPLORATION IT SYSTEMS / DATA MINER / MANAGEMENT VALIDATE STATISTICIAN MODEL TRANSFORM Model Validation & SELECT Exploratory Analysis Model Deployment BUILD Descriptive Segmentation Model Monitoring MODEL Predictive Modeling Data Preparation Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 8. Trends in Platforms $20 $18 Hadoop COST PER TERABYTE $16 COST PER GIGABYTE Microsoft PDW $14 $12 Oracle $10 Greenplum $8 Teradata $6 $4 Vertica $2 $- $- $20,000 $40,000 $60,000 $80,000 $100,000 Today 2009 COST OF STORAGE, MEMORY, COMPUTING In 2000 a GB of Disk $17 today < $0.07 In 2000 a GB of Ram $1800 today < $1 In 2009 a TB of RDBMS was $70K today < $ 20K Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 9. Leverage in-memory architecture via a dedicated software and hardware appliance BIG DATA • Drive high-performance capabilities across the analytical lifecycle MEETS BIG • Achieve insights at breakthrough ANALYTICS speed before questions become PRODUCT HIGHLIGHTS obsolete • Offer a consistent interface for current SAS analytic users Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 10. CREDIT DEFAULT RISK ASSESSMENT APPLICATION SCORING BEHAVIORAL SCORING COLLECTION SCORING RISK ASSESSMENT ANALYTICAL LIFECYCLE Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 11. CUSTOMER TRADITIONAL ANALYTICS PROCESS CASE STUDY 167 Hours DATA MODEL MODEL EXPLORATION DEVELOPMENT DEPLOYMENT Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 12. CUSTOMER HIGH PERFORMANCE ANALYTICS PROCESS 167 Hours CASE STUDY DEVELOPMENT E X P L O R AT I O N DEPLOYMENT MODEL MODEL Bottom-line Impact: D ATA Tens of Millions of Dollars 84 SECONDS Copyright © 2011, SAS Institute Inc. All rights reserved.
  • 13. C op yr i g h t © 2 0 1 2 , S A S I n s t i t u t e I n c . A l l r i g h t s r es er v e d . SAS.com

Hinweis der Redaktion

  1. &quot;Big data&quot; is a popular term generally used to acknowledge the exponential growth, availability and use of information (structured and unstructured). A lot has been written lately on big data trend and how it will become a key basis of competition, innovation, and growth.How does SAS define or view the term “Big Data”?Big data is a relative term (and not an absolute term) - when an organization’s ability to handle, store and analyze data (from a volume, variety and velocity perspective) exceeds its current capacity (i.e. beyond your comfort zone) then it would qualify of having a “big data” problem.
  2. Big Data constitutes: Volumes - Growing volumes of data and how much data need to be processed within a time window Variety - includes structured tables, documents, e-mail, metering data, video, image, audio, stock ticker data, and more. Velocity - How fast data is produced and processed to meet demand. Ability to respond once a problem or opportunity is detected. A data environment can become extreme along any of the dimensions or combination of two or all of them at once. Hence it is important to determine and evaluate “relevant” data to answer the complex set of questions you have before they become obsolete. SAS role here – help determine what is relevant and what is not! In any given situation whether you are looking at pole top transformers, coal fired turbines, … or oil drilling equipment 6000 below sea level, or coupon redemption rate at the local grocery storeBig data in and of itself is not that interesting. Good data management practices is the answer to managing big data. But the only way you can leverage “big data” for valuable insights is by using game changing analytics from SAS.  Opportunity for you to excel in your market … or a Ball and Chain to hold you back if you don’t embrace it effectively.** relatve how the values are relative and vary by customer** share examples of yours: Global Oil and Gas company Marketing analytics Service provider to mfg, cpg, and retailTRANSITION – so how do you thrive in big data … solid process, and leverage the right technology … ANALYTICS, ANALYTICS, ANALYTICSText below is from the Jim Davis analytics video on YouTubeOverwhelming amount of data today. And different types of data -- data structured in databases, and then unstructured like voice, video and text.Some call it the data deluge, other say we are drowning in data. We don’t need to look at it that wayLook at data as opportunityNow we may be comfortable making decisions based on gut feel, but that’s not going to cut it [in the smart grid era]The stakes are much higher now. We’ve got to make decisions based on facts.How do we do that? Easy. Analytics can and should be the differentiator.
  3. 1.0 Fundamental set or types of Analytics – which are core to our business and our analytical applications2.0 Customers use a combination of analytical techniques – for example data mining and text mining.3.0 On the front-end Data Management is important because end users spend lot of time and effort in preparing data for analytics.4.0 On the downstream-end, sharing of analytical insights through easy-to-use visualization/BI tools is important. 5.0 Integrated set of components
  4. SAS High-Performance Analytics is delivered as a pre-configured analytics appliance. It includes analytical capabilities spanning data exploration, modeling and scoring from SAS delivered on either Teradata or EMC Greenplum database appliance to solve complex problems in a highly scalable, distributed environment using in-memory analytics processing. It will let customers develop and deploy analytical models using complete data – not just a subset or aggregate – to get accurate and timely insights and take well-informed decisions. It does not limit analytic professionals to using simplified analytical methods for solving complex problems. Compresses or shrinks the time from model inception to model deployment and derive rapid insights to make well-informed decisions or before the questions become obsolete!SAS High-Performance Analytics will include a select set of procedures from following SAS products: Base SAS, SAS/STAT, SAS/ETS, and SAS Enterprise Miner. A SAS 9.3 client interface manages the submission of high performance enabled problems to the compute grid (appliance) for execution.
  5. Credit risk decisions are sensitive but fundamental to banks and lending institutions. Too much credit exposure can lead to high default rates and charge-offs; not enough often means lost business and revenue. Accurate and timely decisions on: accepting an applicant (application scoring), likelihood of defaults among customers who have already been accepted (behavioral scoring), and likely amount of debt that the lender can expect to recover (collection scoring) can easily differentiate a bank as a leader or a laggard in the market. Early detection of high-risk accounts (i.e. for cards, residential mortgages, commercial loans) is critical to perform targeted intervention, reduce bad-debt and reduce overall losses. The need to understand behavior and credit risk exposure at the “customer” level, across all touch points they will have with the bank and across all life-style changes (i.e. time dimension) puts a toll on analytic professionals. Banks or lending organizations need to build a greater number of segment-specific models for a variety of purposes, calculate the probability of default (PD) (as an example) on the loans they service and determine when and whether the borrower is migrating to a riskier pool. Benefits: SAS High-Performance analytics helps to incorporate large volumes of data with no limits on number of observations and variables or attributes for accurate determination of likelihood of defaults and loss forecasting. In addition, it will allow banks to adjust the historical transition probabilities bases on changes in interest rates (or other macroeconomic factors), and hedge these risks effectively. SAS High-Performance Analytics compresses the entire model lifecycle from days or hours to minutes or seconds. Bank will be able to enhance the credit risks to reflect real-world assumptions and include variables across dimensions. It offers analytic professionals the flexibility to test multiple scenarios or new ideas, use best modeling techniques, and perform model iterations more frequently to accurately and quickly identify risks at individual portfolio level and take targeted actions for bank to stay ahead of the market.
  6. The traditional analytical process is time consuming and inefficient. It simply take long time (days) to finish data preparation/exploration, model development and model deployment steps. More specifically what if you :Have problems that they can’t solve because their data volumes are too big or beyond the capacity of their existing systems.Have too many records in their data or too many attributes or variables that needs to be incorporated in the modeling process. Predictive Analytics need to be applied on more granular level data. For example, modeling churn at the customer level instead of branch, predicting parts failure for an entire product line, or modeling propensity to buy at the account level. Massive variable selection stepis required, which necessitates sorting through thousands of variables to determine which are the most predictive. Do not want to compromise by using sub-optimal modeling techniques.Cannot quickly test or experiment different modeling techniques and find the best fit to improve accuracy.Fail in terms of getting desired modeling results and want to modify the model with new attributes and do not have time to wait.