Big Data Tida

  • View
    216

  • Download
    3

Embed Size (px)

Text of Big Data Tida

  • 1 Copyright 2012, Oracle and/or its affiliates. All rights reserved.

    Sustainable Competitive Advantage Through Better Decision:

    How Big Data Information Discovery Provides Valuable Insights

    Tidaporn Santimanawong, Sales Consulting Director, Oracle ASEAN Business Analytics

  • Business Intelligence

    (Multi-dimensional)

    2 | 2012 Oracle Corporation | Confidential Oracle Internal

    (Life Style, Value)

  • Generation Y, Young Generation

    Generation Y 35

    3 | 2012 Oracle Corporation | Confidential Oracle Internal

    55% (1,024 )

    42%

  • Generation Y, Young Generation

    Y

    4 | 2012 Oracle Corporation | Confidential Oracle Internal

    + ,

  • Oracle Business AnalyticsDrive Better Business Outcomes

    ANY

    BETTERDECISIONS

    Copyright 2012, Oracle and/or its affiliates. All rights reserved. Oracle Confidential.5

    ANYDATA

    FASTERACTION

  • Why should I pay a checking fee if other

    banks dont charge

    one? customer.

    She saved the project

    with strong leadership

    & building trust with the

    customer.

    Dialogues Contain Critical, Untapped Insights

    discounted shipping.

    Competitive pricing is

    15% lower than yours

    and they offer

    discounted shipping. late

    Refuse collection hasnt been so reliable recently 8 theyve been coming

    late

    Copyright 2012, Oracle and/or its affiliates. All rights reserved.6

    Customers Workforce Supplier &Partners

    The Public

  • New Insights Drive New Opportunities

    Revenue & margin Customer retention Brand perception Segmentation and

    targeting

    Impact

    Employee retention Improved

    productivity Hiring& Staffing Compensation

    Customers Workforce

    Copyright 2012, Oracle and/or its affiliates. All rights reserved.7

    Revenue & marginOperational efficiency.Better partnershipsProduct positioning.

    Impact

    Revenue & margin Operational efficiency Better partnerships Product positioning

    Public sentiment Trends insight Brand perception Hiring ability

    The Public3rd Parties

  • The Challenges of Unstructured Data

    DATA IS GROWINGIN VOLUMEAND DIVERSITY

    DATA CAN BE DIRTY OROF UNCERTAIN VALUE

    XML

    AMOSTLY TEXT, ANDDIVERSE SCHEMAS

    Copyright 2012, Oracle and/or its affiliates. All rights reserved.8

    Websites Social MediaText in Enterprise

    Applications

    Enterprise Content Systems,

    File Systems, EmailBig Data

    80% UNSTRUCTURED

    Business Intelligence

    and Data Warehouses

    20% STRUCTURED

  • Oracle Endeca Information DiscoveryBest platform for Unstructured Analytics

    Endeca ServerHybrid Search/Analytical Database

    Flexible Data Model

    Oracle Endeca Information DiscoveryExtend Business Analytics with Unstructured Data

    Oracle Business IntelligenceBest platform for integrated ROLAP and MOLAP

    BI Server + EssbaseCommon EnterpriseInformation Model

    Copyright 2012, Oracle and/or its affiliates. All rights reserved.9

    Social MediaContent Systems,Files, Email

    Websites

    Unstructured Data

    Big DataOLTP & ODS

    SystemsEnterprise Applications(Oracle, SAP, Others)

    Data Warehouse& Data Marts

    Structured Data

  • High Value Information Discovery use cases

    impossible to address with traditional BI tools

    Four example use cases:

    1. Product Quality - Discovery and Analysis :- ERP Manufacturing Apps

    2. Insurance Claims - Discovery and Analysis :- Insurance and Healthcare customers

    10 | 2011 Oracle Corporation | Confidential Oracle Internal

    3. Criminal Intelligence - Discovery and Analysis :-Law enforcement and intelligence

    agencies

    4. Social Media and Sentiment - Discovery and Analysis:- CRM Application

  • Use Case 1: Product Quality Discovery & Analysis

    CUSTOMER PROFILE

    Discrete manufacturer

    Product population

  • Reduce warranty costs and detection to correction time

    Common view of six years of product and quality data from

    Investigate Root Cause

    Global Automotive Manufacturer Gets Ahead of Quality Issues

    Copyright 2012, Oracle and/or its affiliates. All rights reserved.12

    product and quality data from multiple systems

    Quality engineers explore root cause (why) of product issues from unstructured data.

    Operates along side Oracle BI EE, Essbase, EPM and PeopleSoft

  • An Example: Automotive Quality

    Suspect

    Algorithm

    Correlation Analysis

    identifies a high rate

    of battery failures in

    Scandinavia

    Using Endeca the analyst The Analyst Discovers:

    13 Copyright 2012, Oracle and/or its affiliates. All rights reserved.

    ExploreDiscover

    Data:

    Structured &

    Unstructured

    Using Endeca the analyst

    investigates: Structured Data

    - Defective parts and suppliers

    - Vehicle models and options

    - Weather patterns

    - Manufacturing sites and dates

    - Vehicle distribution and delivery

    Unstructured Data

    - Service Technician Reports

    - Internet Forums

    The Analyst Discovers: Service technicians and vehicle owners

    have complained about the problem in unstructured text - and also mentioned seat heating

    The alternator on problem vehicles has recently been second sourced from a new supplier

    This new alternator ...

    ... delivers insufficient power

    ... in extreme conditions

    ... in vehicles that have seat heating

  • Opportunity 2: Insurance Claims Discovery & AnalysisWho buys it & what are they looking for?

    CUSTOMER PROFILE

    Insurance provider

    Active claims between 1M and 10M

    CUSTOMER PROFILE

    Insurance provider

    Active claims between 1M and 10M

    PAINS

    Claims management system notes cant be incorporated into analysis because it is in unstructured (text) format

    Difficult to identify fraudulent claims without ability to correlate information between different

    PAINS

    Claims management system notes cant be incorporated into analysis because it is in unstructured (text) format

    Difficult to identify fraudulent claims without ability to correlate information between different

    OBJECTIVES

    Earlier, more accurate detection of fraudulent insurance claims

    Unlimited exploration across all dimensions of the source data systems

    OBJECTIVES

    Earlier, more accurate detection of fraudulent insurance claims

    Unlimited exploration across all dimensions of the source data systems

    14 | 2011 Oracle Corporation | Confidential Oracle Internal

    information between different sources

    Existing BI tools arent designed for use by front line claims analysts with domain expertise

    information between different sources

    Existing BI tools arent designed for use by front line claims analysts with domain expertise

    Consumer-friendly, search-oriented interface

    Reduced overall claims cost and increased customer satisfaction

    Consumer-friendly, search-oriented interface

    Reduced overall claims cost and increased customer satisfaction

    The notes in our claims management system

    are finally available alongside our structured

    data sources, supporting our fraud detection

    efforts and improving call handling quality.

    CIO

  • Customer Success: Liberty Mutual

    Auto, Home and Life Insurance Provider

    OVERVIEW

    Liberty Mutual needed to reduce claims fraud and abuse while improving overall customer satisfaction

    CHALLENGES/OPPORTUNITIES

    Existing BI standard (MicroStrategy) unable to combine notes fields from insurance claims with structured data sources

    RESULTS := Go-live 10/21/2011

    15 Copyright 2012, Oracle and/or its affiliates. All rights reserved.

    insurance claims with structured data sources Competitive pressure from low-cost providers to differentiate by providing best-

    in-class customer service

    SOLUTION

    Provides a set of combined views across structured and unstructured data from over 10 million active personal market claims

    360 degree view of each claim enables faster and more accurate identification of fraudulent and abusive claims

    Front line claims analysts can access data while on the phone with a customer to improve customer responsiveness and satisfaction

  • Use Case 3: Criminal Intelligence Discovery & Analysis

    CUSTOMER PROFILE

    Law enforcement agencies

    Intelligence agencies

    City, state and national government

    CUSTOMER PROFILE

    Law enforcement agencies

    Intelligence agencies

    City, state and national government

    PAINS

    Information can be useless by the time it is cleaned, processed and loaded into a warehouse

    Reduction in operating budget and operations staff while activity levels are increasing limits ability to quickly perform crimes analysis

    Activity history spread across several

    PAINS

    Information can be useless by the time it is cleaned, processed and loaded into a warehouse

    Reduction in operating budget and operations staff while activity levels are increasing limits ability to quickly perform crimes analysis

    Activity history spread across several

    OBJECTIVES

    Advance detection and crime prevention

    Reduce operating costs and increase intelligence visibility

    Efficient deployment within first line o