Data Warehousing / Business Intelligence

صفحه 1:
۱1: 7 ۱۵2۵ Business Intelligence IT Solutions. Guaranteed.” www.eforceglobal.com 

صفحه 2:
Outline ‏شلال‎ ‏جوصونص 8) ا"‎ Toteliqeuse — Business Ieperaive, Gtrateqic Drivers = OO/@1 - Provess 2 OO/B1 - ‏)اوه‎ Cxperise ® Ouse Grdy ® Ophodobow/Oevebpwedt Process 

صفحه 3:
Business Case for Bl 0۱۳ Competitive Advantage Predictive Metrics “What will happen?” Descriptive Metrics “Why has it happened?” Complexity of Analysis Higher Baseline Metrics “What has happened?” Higher Business Value Recommended ‘Actions 

صفحه 4:
Data Warehousing 0۱۳ #۱ | oa ۷ لمسیره 

صفحه 5:
A subject oriented approach ۱۳۹۳ Operational Systems Data Warehouse Ores a. ۵۰۳۰۳۰ Owtewer Googly و ‎Product‏ Organized by processes Organized by subject or tasks 

صفحه 6:
Data Warehouse Process 0۱۳ 0 9 Gece 0 ]1990 این ‎Dexter,‏ قاس 9 سس ‎[Bdge‏ 

صفحه 7:
BI Reference Architecture 0۱۳ تس ‎CR RCE la)‏ لل ‎of Business Intelligence ‎ ‎ 

صفحه 8:
BI Reference Architecture 0۱۳ 

صفحه 9:
BI Trends 0۱۳ 8 dooreuse ‏امه ع ۲ ول ۲ وتا مود و‎ ico OO BED ‏بلج‎ )2٩ ‏امه‎ ‎۲ ‏رررشهظ) عویمنی()‎ )(0۰۲۲ ‎ooreuse urderstantiog oF Oetadata ood Ispportcace i (B1‏ ا" ۳۲۳۳ ‎™ Deb Genvices ‎® Data Quatiy “Pirewls” ‎ 

صفحه 10:
Skill Sets ۳ 6 شمه LOLOP Tooke | ‎mene eevee‏ ا ‎Cees es eh eek eed‏ ‎ 

صفحه 11:
DW driven - Customer Value Management ۱۰ Apply the customer insight, at the moment and point of Accumulate customer and customer contact, using the transaction data from multiple latest available information, to GATHER: sources 9 personalize the interactions 6۶ 3 APPLY ‏و‎ g ANALYZE \ x Use incight to etine’a campaign ۹ Analyze data (data mining) to according to a marketer's develop a fact based and objectives, either in real time for statistically sound ‘online channels or in batch for understanding of customers offline channels and prediction of their needs and behavior 

صفحه 12:
1۱1۱: Customer Value Management Context Unified Customer View (Joe's Real-Time Data and Context-Defined Profile| 3 3 8 ع 2 ‘Account (IRA) ولمع عبااه/ا 27 ‎Ag Income‏ وغ عبادلا ‎rs As Cust ‘Amount x 12‏ لامدممروعء ‎Probability‏ 4% و عبادلا ‎Creston of Response‏ ‎۹ eo Recommendation NG ‎. ‎8 ‎۴ ‎9 ‎a ‏ع‎ ‎2 ‎2 ‎8 ‎8 ‎2 ‎Gg ‏هه‎ ‎6 ‎Real-Tim ‎Recommendation ‎to Xsell Credit Card ‎ 

صفحه 13:
Case Study - Real Time Business Intelligence 

صفحه 14:
Case Study - Real Time Business Intelligence - ۱۰ اه ‎٩‏ مره مرو ‎Por‏ ای اواج و وم عصو ول 1 ]0 ‏مرن امه سوت 003 يعنص لوا‎ KG] (PerPors the racks oP brody updated dota Prow pxieronl syoieore, ore aggre ond derived vdeo moxk ke OR, od oreutey dra exports Por ver ‏مسد‎ ] ‏سس‎ ( (Prous fe primary wank nto the dra. (Deed Por reporters ‏,ماد ,سای تسا لی‎ ] acre ver they crest set oF took: (6DG, (BPR axRo Dede, et) Por exploratory samo, wen oreaiva or ‏رای جات‎ propos deci Prow ie (BUR cha store. Denne te proxkctva versie oP exniic woodeks thot rcripks ces waka he (BAR by ‏له وه مره عه تسه مان افو اوه مططلی:‎ predeave purchase sours. ‘Nhe cubkious ue perPorwed by profile avout ‏امه رفسمه مج‎ Grace Dvexess Doubse (rou. Dik he profle cyputs wocaced ia a couse Packs, ‏تا طاط سا بت سای‎ reuse post cords raher koa ore GPU OWES. [Chern developed vor COG barter] “Dhe core dara store oF the (B10 taker rerercrtva level dha ‏لجيج جنب جا دجت لحت‎ Bosdrae roped corres Por 1 vere. Dhewr arate are cxngrected by axkdioand sorts tt exten a vaey oP aepreqaied, puDarted, out derived vole ceeded Por caayical pecans Components ved Loader | Derr ane OT) Phforw عمط وه Bede Toot Support PrP Bayer Ons Gore 

صفحه 15:
Case Study - Real Time Business ۱۰ Intelligence - Benefits ‎wak Ord pay COG took: wo prove Blexble compar‏ موی ‎bur O11 eek‏ يحاص مومس ‎,] oom be everced whether we ue our ODO or te chests! ‎(Reduced heey oP nPoreraioa updater by cove beyoed bark proceroka| 1 wear red ioe [red ‏سيد لت سلجن مي‎ ‎lhoproved ieveteese oF abst ced rewara kre Keadhog io Roproved ‏مه اس مها | مهو‎ ‎CP roPe expe represeuten 9 “sioraxp box” oP corde (Proprietary statcard wodehy | predietve smpoeciaioa cupdltior ho ts PE ‎i code‏ سول ‎creche i prove‏ مه بو یه تا ما ‎(01 cata pushed frocks wale chases erkudiog wal, eco, web, ‎touch ports, oe (Buen detects creas Por highhto hi ‏مخ‎ oan Por igri toa. ‏ا حوطس سوام ممم ایا ماه با وم‎ ‏ای ,و جوا ‎share key‏ 1 قدت افك ايف ‎pron ave ord‏ ‎ress, Pe‏ ون هو ‎ ‎ ‎ ‎eter tteqrutza wis oie sews ‎ ‎ ‎ ‎ ‎ ‎ ‎ ‎ 

صفحه 16:
Development Methodology 0۹ 

صفحه 17:
Data Warehouse Life Cycle Model ‏شلال‎ ‎a=) ‎- ‏مسج لا‎ ‏سوه‎ ‎Ad [ 4 > Oevebp Ores ۱ oom se ‏سین باه‎ اي 

صفحه 18:
insite 

صفحه 19:
Execution Methodology 1۱۰ 

صفحه 20:
Maintenance Approach ۱: 0: 

صفحه 21:
سرت High-Level Project Plan 8 ‏علد يسما يسع‎ ete 7 

صفحه 22:
۱1: Summary ۷ ‏و6‎ OO, ©1 experiewe Proves OO/O1 Cbd Oebery Dodel Cxperter to best to ches eaublay techoby panes ۷ Guperer یت ‏مم مرت هت‎ 

صفحه 23:
7 Force IT Solutions. Guaranteed.” صمءع www.eforcegloba 

Data Warehousing/ Business Intelligence IT Solutions. Guaranteed. TM w w w . e f o r c e g l o b a l . c om Outline  Business Intelligence – Business Imperative, Strategic Drivers  DW/BI - Process  DW/BI - Technology/Package Expertise  Case Study  Methodology/Development Process eFORCE Proprietary & Confidential 2 Business Case for BI Competitive Advantage Higher Predictive Metrics “What will happen?” Descriptive Metrics Business Value “Why has it happened?” Baseline Metrics “What has happened?” Lower Higher Complexity of Analysis Data Information Analysis eFORCE Proprietary & Confidential Insight Recommended Actions 3 Data Warehousing Information Operational data eFORCE Proprietary & Confidential Knowledge workers 4 A subject oriented approach Operational Systems Data Warehouse Sales Customer Financial Geography Inventory Organized by processes or tasks Product Organized by subject eFORCE Proprietary & Confidential 5 Data Warehouse Process DATA SOURCES STAGING AREA DATA WAREHOUSE Application Databases DECISION SUPPO _________ _________ _________ _________ _________ _________ _________ _________ _________ _________ _________ _________ Reports Packaged application/ERP Data DATA MARTS INCOME INCOME ANNUAL ANNUAL REPORT REPORT ___ ___ ___ ___ ____ ____ _____ _____ ___ ___ __ __ ___ ___ ___ ___ ____ ____ _____ _____ ___ ___ __ __ ___ ___ ___ ___ ____ ____ _____ _____ ___ ___ __ __ Desktop Data EXTRACTION TRANSFORMING CLEANING AGGREGATION EIS DATA WAREHOUSE OLAP External Data Statistical & Financial Analysis Web-based Data [Adapted from SunExpert Magazine, October 1998.] eFORCE Proprietary & Confidential 6 BI Reference Architecture - Gartner eFORCE Proprietary & Confidential April 2004 7 BI Reference Architecture eFORCE Proprietary & Confidential 8 BI Trends  Increase  ETL in zero latency for data to be available in a DW and AI technologies merging  Business Activity Monitoring  Increase understanding of Metadata and Importance in BI projects  Web Services  Data Quality “Firewalls” eFORCE Proprietary & Confidential 9 Skill Sets Databases Oracle, Sybase, SQL Server, Informix, Red Brick ETL Tools Informatica, Ascential Datastage ROLAP Tools Microstrategy, Information Advantage, Informix Metacube MOLAP Tools Cognos, Hyperion Essbase, Oracle express, Brio, Business Objects HOLAP Tools Seagate Info, Holos Web-enabling Tools Cold Fusion 4.x, Microstrategy DSS Web, Brio On demand server, Cognos Powerplay Web, Business Objects- web eFORCE Proprietary & Confidential 10 DW driven Customer Value Management Apply the customer insight, at the moment and point of customer contact, using the latest available information, to personalize the interactions GATHER APPLY Use insight to define a campaign according to a marketer’s objectives, either in real time for online channels or in batch for offline channels Accumulate customer and transaction data from multiple sources ANALYZE MODEL eFORCE Proprietary & Confidential Analyze data (data mining) to develop a fact based and statistically sound understanding of customers and prediction of their needs and behavior 11 Customer Value Management Joe Unified Customer View Context “Amount” DD/Interaction Server Joe Real-Time Recommendation to Xsell Credit Card (Joe’s Real-Time Data and Context-Defined Profile) Amount=$1400, Age=46, Income=… Qualify Available Business Objectives X Xsell Savings Account Xsell Credit Card (CC) Xsell Retirement Account (IRA) Dynamic Value Prioritization Value to Company Age Years As Cust Income Amount x 12 Value to Customer 32.5% Prob of Response 9.4% Probability of Response Overall Value of Recommendation ($18)x(0.33) = $5.98 ($5.1)x(0.94) = $0.48 X OLTP Case Study – Real Time Business Intelligence ECIS Campaign Mgmt AIR Message Queue Request Response Messages Event Handler ETL Platform Bulk Transaction Updates Profile Engine Data Store Aggregate Core Segmentation/ Communication Events Business Objects U U Report U Analytic Tools U Report SAS Model 1 eFORCE Proprietary & Confidential 13 Case Study – Real Time Business Intelligence Components Event Handler / Message Message queue provides a standards-based mechanism for application integration. The event handler ties messages to the appropriate containers Queue [Deployed using MQ Series and Open for JMS] ETL Platform Performs the tasks of loading updated data from external systems, creating aggregates and derived values inside the AIR, and creating data exports for other systems. [Informatica 7.1] Business Objects Provides the primary window into the data. Used for reporting, ad hoc analysis, dashboards, etc… Analytic Tool Support Allows analysts to use their current set of tools (SAS, Affinium Model, etc) for exploratory analysis, model creation or other analytic projects directly from the AIR data store. Profile Engine Manages the production versions of analytic models that enrich data within the AIR by calculating complex derived values such as segmentation assignments and predictive purchase scores. The calculations are performed by profile agents that are created by Targetbase’s Strategic Business Analysis group. With the profile agents managed in a consistent fashion, analysts will be able to reuse past models rather than creating new ones. [Custom developed using CVS backend] Data Store The core data store of the AIR takes transaction level data and stores it in marts organized to facilitate rapid access for BI uses. These marts are augmented by additional marts that contain a variety of aggregated, summarized, and derived values needed for analytical operations. eFORCE Proprietary & Confidential 14 Case Study – Real Time Business Intelligence - Benefits  Integration with 3rd party CMS tools to provide flexible campaign management using our BI methods  BI analytics can be leveraged whether we use our CMS or the clients’  Reduced latency of information updates by moving beyond batch processing to near real time / real time updating and scoring  Improved timeliness of analysis and reaction time leading to improved interactions / relationships with customer Automation of our key analytical methods - Profile Engine  Profile engine representing a “storage box” of models  Proprietary statistical modeling / predictive segmentation capabilities live in PE  “Storage box” acts as services mechanism to provide efficiencies in model execution Expanded channel integration capabilities  BI action pushed through multiple channels including mail, email, web, operational touch points, etc. Alerts and Detects  Event detection methods for highlighting latent findings and for triggering action Improved user experience  Facilitates the attachment of web-based information dashboards designed for executive and non-technical entities to share key insights, trends, actionable metrics, program results, etc. Better integration with other systems Improved throughput eFORCE Proprietary & Confidential 15 Development Methodology Business Requirements Planning Using RAD Method Conceptualization Using JRP Sessions Business Requirements Definition Using JAD Sessions Design & Prototyping Architecture Construction Testing 30- 35 % 60 - 65 % eFORCE Proprietary & Confidential 16 Data Warehouse Life Cycle Model Software Concept Preliminary Requirement Analysis Deliver Final Version Architecture Develop Version Deliver Version1 Incorporate Customer Feedback Version Version 2 …n Elicit Customer Feedback eFORCE Proprietary & Confidential 17 Development Methodology Business Requirements Planning Onsite Conceptualization Off shore Business Requirements Definition Integration & System Testing Design & Prototyping Architecture Construction Implementation Unit Testing eFORCE Proprietary & Confidential 18 Execution Methodology Production Onsite Requirements Gathering Iterative Development Analysis / Prototyping Acceptance Testing High Level Design Offshore High Level Design Detailed Design eFORCE Proprietary & Confidential Construction Testing Integral / System 19 Maintenance Approach Ongoing Onsite/Offshore Onsite/Offshore Support Support 5 Weeks Application Understanding Understanding Maintenance Process Study Start Support Activity Offshore Offshore Transition Transition Planning Planning Requirements Study Transition Planning Cost Estimation Account Account Manager Project Manager On-site Team Team eFORCE Tools Technical Environment Study PETE/IAN Knowledge Repository QA QA & & Performance Performance Monitoring Monitoring Value Added Services 3 Weeks Planning Planning Knowledge Transfer Knowledge Transfer Service Service Coordination Off-shore Off-shore Team eFORCE Proprietary & Confidential 2 High-Level Project Plan Inception: 6 Weeks Activities       Detailed mining, assessment and re-use of existing deliverables Vision and critical success factor development Shared view between team and stakeholders on scope and direction Activities   Detailed requirements analysis Some key offshore personnel remain onsite, others go home to startup offshore team   Complete system architecture and detailed design  Proof of concepts/initial development for key technical areas (CCI integration, CTI integration, XML integration, MU integration, RSL integration, security, etc.)   Establishment of the Change Control Board and associated process Development and Unit Test Production Rollout Daily Builds  Technology Transfer Weekly Integration Builds  Training  Frequent Code Reviews   Quality Assurance Gradual Phase-Out of Legacy Code Base   Unit Testing System Testing  Integration Testing  User Acceptance Testing Architectural review and certification  Performance Testing Deliverables  Vision  Requirements Documents (Detailed)  Critical Success Factors  Architecture Certification  Master Test Plan   Architecture Certification Plan  Product Transition Plan Design Deliverables (Software Architecture Document, Design Models, Site Navigation Map, Wireframes, Design Guidelines, ClickThrough Site Mockup) Usability Test Results  Documentation Plan   Communications Plan  Portal Integration Strategy  Task Plan/Resource Plan  Test Cases  Information Security Plan  Knowledge Transfer Plan  Requirements Documents Activities  Progress being made on distributed testing strategies Deliverables Transition: 6 Weeks Activities   Key offshore personnel brought onsite Identification of, and initial interactions with, stakeholders and SMEs (CCI, RSL, APS, ECPR, ODS/PAS, TQ, VL, etc.) Construction: 20 Weeks Elaboration: 8 Weeks Deliverables  Build Deliverables (Application Build, Installation Artifacts, etc.)  Testing Deliverables (Test Cases and Test Results)  Code Base  Database Scripts  Deployment Plan eFORCE Proprietary & Confidential Deliverables  iBK Architecture, Design, Development, and Test Handbooks for Knowledge Transfer  iBK Architecture, Design, Development, and Test Customized Training Workshops  System Capacity Model and Performance Test Results  Release Notes  Level 3 Support 21 Summary  Superior DW, BI experience   Proven DW/BI Global Delivery Model   Expertise in best in class enabling technology packages   Superior implementation methodology  eFORCE Proprietary & Confidential 2 IT Solutions. Guaranteed. TM w w w . e f o r c e g l o b a l . c om