1 Components of A Successful Data Warehouse Chris Wheaton, Co-Founder, Client Advocate

1

Components of A Successful Data Warehouse

Chris Wheaton,

Co-Founder, Client Advocate

2

Presentation Information

• Presentation: Successful Components of a Data Warehouse

– The purpose of this presentation is to provide attendees with the understanding of how to build a successful data warehouse/business intelligence solution.

3

Presentation Information

• Author: Chris Wheaton

– Biography: Chris Wheaton is a Co-Founder of BASE Consulting Group, Inc. He initiated and has contributed to the development of the Business Intelligence and Data Warehousing Certificate Program at the University of California, Berkeley Extension and is a lecturer in the program. He has presented on Data Warehousing topics at conferences throughout the United States including the Business Objects and Oracle Applications User Conferences.

– Contact Information:• E-mail: [email protected]

• Phone: (510) 628-3300 Ext. 223

4

Agenda• Data Warehouse ? What is it? Why do it?• Why do they fail?• How do you control business risk?• How do you control technical risk?

5

Many names for the same thing

• Decision Support System (DSS)• Executive Information System• Management Information System• Business Intelligence Solution• Analytic Application• Data Warehouse

Six terms for the same thing: a system for helping companies get the information they want, when they want it…etc.

Data Warehouse? What is it?

6

Data Warehouse Objectives

- Business• Access to specific high-value information on a timely basis.

Analysis• Makes the restructured data available to users via user-

friendlier query and reporting tools. Reporting

- Technical• Gets the data off of the transaction system for analysis.

Performance• Restructures and integrates the data so that it is easier to use

for reporting and analysis. Integration

Data Warehouse? What is it?

7

Data Warehouse Risks

Business• Content

– Does the solution answer the right questions?

– Does the solution have enough data?

• User Acceptance– Is it too complex for the

average user?– Is the data timely enough?

Technical• Performance

– Is the performance of the user queries satisfactory?

– Can data be loaded to the data warehouse within the allotted timeframe?

• Integration– How do we combine

information from multiple systems?

Why do they fail?

8

Controlling Business RiskWe have found that the best way to address the business risks associated with a data warehouse project is to employ a methodology with the following components:

– Enterprise Strategy

– Phased Delivery

– Iterative Prototyping

How do you control the risks?

9

BASE Methodology

PROJECT MANAGEMENTCLIENT ADVOCACY

SoftwareSelection

Inform ationAccess

Roadm apPresent

Prototype

FunctionalSpec

Design

Build

SustainingOperations

Transition

DeployTrain

31

4

62 87

5

9.1

9.2

Strategic Tactical Operational


10

Enterprise Strategy• Asses technical landscape• Identify business drivers• Define analytical processes to be supported• Identify major facts, dimensions and attributes• Map and gap to data sources• Assess current architecture and tools• Recommend subject area phasing and tool

selection


11

Phased Subject Areas


____________________________1 2 3 7

Conceptual Architecture

SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Phase 1Prototype 1

Phase 1Requirements

Phase 2Prototype 1

11

Transition

4

Phase 1Prototype 2

5

Phase 1Prototype 3

Phase 1 Deployed to Production

6

Deliverable #

Milestone

8

Phase 2Prototype 2

9

Phase 2Prototype 3

Phase 2 Deployed to Production

10

Phase 2Requirements

12

Subject Area Scope

Subject Area Focus

• A related set of business

entities as defined by specific

user group.

• Integration between subject

areas.

Examples are:

• Bookings, Billings, Backlog

• Customer Support Calls• Inventory • Marketing • G/L Transactions


13

Iterative Prototyping

IdentifyUsers

DeliverPrototype

DevelopDesign

SelectProtoype

DesignElem ents

BuildPrototype

MigratePrototype toProduction

Docum entRequirem ents

ConductInterview s

SelectSubject Area


14

Controlling Technical Risk• We have found that the best way to address the technical

risks associated with a data warehouse project is to employ an architecture with the following components:– Integrated Staging Area

– Dimensional Data Store

15

Architectural OverviewSource Data

Staging

Staging Area DimensionalData Store

End User Reportingand Analysis

General Ledger(GL)

Fixed Asset (FA)

AuthorizedExpenditures

(AFE)

Staging

Staging

Transform& Load

Query

Staging Database

StagingFiles

Meta Data Repositories

Reporting Tool

Tax Information

AFE Information

FA Information

GL Information

Relational DataWarehouse

(Oracle RDBMS)

Data Warehouse Hardware and IS Development & Maintenance Tools(Sun E450, Informatica, Oracle DBMS Utilities)

Reconcile

Query

Data Warehouse – What do you deliver?

16

Integrated Staging Area

• Database tables holding production data before it is loaded

into the dimensional data store tables.

• Tables usually resemble the source data tables, and have

not been re-structured except to allow for some

integration.

• Critical to source system reconciliation.

• Staging area may also include flat files in original format.


17

Dimensional Data Store

Sales by Product &

Day

Sales

Time Geo

Product

Aggregates

Sales by Geo &

Day

Dimension

Tables

Fact Table


18

Advantages of a Dimensional Model• Standardization of dimensions helps standardize reporting

across areas of the business.

• Dimension tables preserve the history of the dimensional information.

• Whole new dimensions can be introduced without major disruptions to the fact table.


19

Extract, Transform and LoadSource Data

Staging

Staging Area DimensionalData Store

End User Reportingand Analysis

General Ledger(GL)

Fixed Asset (FA)

AuthorizedExpenditures

(AFE)

Staging

Staging

Transform& Load

Query

Staging Database

StagingFiles

Meta Data Repositories

Reporting Tool

Tax Information

AFE Information

FA Information

GL Information

Relational DataWarehouse

(Oracle RDBMS)

Data Warehouse Hardware and IS Development & Maintenance Tools(Sun E450, Informatica, Oracle DBMS Utilities)

Reconcile

Staging Tools Cleansing Tools

TransformationTools


Documents

1 Components of A Successful Data Warehouse Chris Wheaton, Co-Founder, Client Advocate