11 Data Warehouse TransUnion Ltd

							Data Warehouse – TransUnion Ltd




                                  © 2006 TransUnion Limited
                                         All Rights Reserved
 Agenda


• A brief of our business background in HK

• Business growth priority

• Steps taken and challenge encountered

• Result obtained and benefit gained in our
  business

• The technology we relayed on for our success


                                                 2
TransUnion HK Background


• Established in 1982 as Credit Information Services Limited
  (CIS) by 12 founding shareholders from Finance Houses
• Curbed serious losses from multiple finance
• Centralized database by 3rd party to share collateral
  information under finance
• Data Collection
   – Started with Hire Purchase files
   – Mid 80s - collected negative files
   – Late 80s - collected public records
   – Aug 2003 – shared Positive Data



                                                         3
TransUnion HK Background            (Cont’d)



 • To be a world-class fully-fledged credit bureau, TransUnion
   acquired a majority stake and became the largest shareholder of
   CIS in 1999

 • Soon after the acquisition, TransUnion brought into the marketplace
   a new credit report system platform “2000Plus”

 • Officially changed name to “TransUnion Information Services Ltd
   (TUIS)” in Jan 2003 and to “TransUnion Limited” in 2006
   (“TransUnion HK”)

 • TransUnion HK is now playing a pivotal role in the financial
   services industry in Hong Kong


                                                                  4
Our Shareholders

• TransUnion HK is privately owned by 8 shareholders
 - TransUnion, D&B and the following banks:

  American Express Int’l., Inc.     The Bank of East Asia, Ltd.


   Hang Seng Finance Ltd.              DBS Bank (HK) Ltd.


 The Hongkong and Shanghai        Standard Chartered Bank (HK) Ltd.
  Banking Corporation Ltd.




                                                                  5
TransUnion Members in HK


• Exclusive Membership under the Privacy Code

  – Retail Banks

  – Credit Card Issuers

  – Bank-affiliated Finance Houses

  – Other Financial Institutions




                                                6
Portfolio Lines in TransUnion HK


      Unsecured Lending                    Secured Lending
•   Credit Cards/Debit Cards       •   Mortgage
     – Private label, affiliate,       (Negative Only)
       co-branded                  •   Mortgage Overdrafts
•   Personal Loan                      (Negative Only)
•   Revolving/Term loans           •   Autos
•   Tax Loan                       •   Vessel
•   Overdrafts                     •   Equipment
•   Charge cards                   •   Engine




                                                             7
Business Growth Priority

• Need to diversify the current static core business
  service (Credit Report)
• Turn data analysis service into business
  opportunities
• Time to market analysis service
• New products are required to develop
 – Credit Scores
 – Customized MIS
 – Market trend monitoring
   ▪ Understand the market and the customers



                                                       8
Steps taken and challenge encountered



• Information locked in business silos and
  applications
• Redundant data, applications and infrastructure
• Can’t deliver information at users need
• Lack of historical snap shot image
• No single view of the customer
• High cost to address each of the above issues




                                                    9
  Result obtained and benefit gained in our
  business

                      Pre                                                  Post
•Scope of ad-hoc data study confined to current   •Historical database up to 24 months always
month plus one previous month data only           available for all kinds of data study

•Require at least 1-2 days to reload historical   •Either ad hoc data analysis or in plan statistic
database before any retrospective data study      process can be performed instantly – time
can be performed

•Upon priority schedule with occupying the host   •Independent system resources for data analysis
server resource and interrupt the scheduled       and products development with real time access
process work

•Inconsistent historical data structure and       •Consistent data structure standard with data
compatibility due to system enhancement           patching control

•Multiple data sources located in different       •Consolidated data pool with one stop data shop
applications platforms

•External data analysis tool is required          •Basic Data Mining and OLAP are easy to executed

•Long time required for new analytic product      •Over 40% of new analytic product development
development                                       cost is saved



                                                                                                  10
     Data Warehouse

                              Depends on the
                              update frequency;
                              can be daily or
                              weekly, monthly, …

                                                                    Ad-hoc
                                                                    Image
   Other
applications             Ad-hoc image
    DB
        TUCRS                 Transformation            Data
      (Informix)                                    Warehouse
                         Month-end image
                                                   Database (DB2)
                                                                    Historical
                                                                    Archive
 Production Database
                                                                    Images
 With up-to-date image




                                                                                 11
    The technology we relayed on for our
    success

•       Low cost hardware platform
    –     RedHat Linux on Intel Xeon CPUs
    –     64G memory
    –     2.5TB Enterprise Virtual Array (EVA) disk storage
•       IBM data warehouse enterprise edition V9.1.2
    –     DB2 enterprise server edition 9.1
    –     WebSphere Application Server V6.0
    –     Alphablox 8.4
•       Performance features
    –     Table Partitioning
    –     MQT (Materialized Query Table)
    –     MDC (Multi Dimensional Clustering)
•       Data volume handling
    –     1.5TB database size
    –     125M credit data transactions during month end transformation
•       Next step in plan
    –     Port to a 64bits environment to gain a better performance

                                                                          12
13
14