TDWI DW Concepts and Principles Preview

Previews of TDWI course books are provided as an opportunity to see the quality of our material and help you to select the courses that best fit your needs. The previews can not be printed. TDWI strives to provide course books that are content-rich and that serve as useful reference documents after a class has ended. This preview shows selected pages that are representative of the entire course book. The pages shown are not consecutive. The page numbers as they appear in the actual course material are shown at the bottom of each page. All table-of-contents pages are included to illustrate all of the topics covered by a course.

TDWI Data Warehousing Concepts and Principles

© The Data Warehousing Institute i

TDWI Data Warehousing Concepts & Principles an Introduction to the Field of Data Warehousing


ii © The Data Warehousing Institute

The Data Warehousing Institute takes pride in the educational soundness and technical accuracy of all of our courses. Please give us your comments – we’d like to hear from you. Address your feedback to:

email: [email protected] Publication Date: May 2004

© Copyright 2002-2004 by The Data Warehousing Institute. All rights reserved. No part of this document may be reproduced in any form, or by any means, without written permission from The Data Warehousing Institute.


© The Data Warehousing Institute iii

Module 1 Data Warehousing Concepts …………................ 1-1

Module 2 Data Warehousing Architecture ………...…….... 2-1

Module 3 Data Warehouse Implementation ….....……….... 3-1

Module 4 Data Warehouse Operation …………………...…. 4-1

Module 5 Summary and Conclusions …….....…..……….... 5-1

Appendix A Bibliography and References …………………… A-1

TAB

LE O

F C

ON

TEN

TS

TDWI Data Warehousing Concepts and Principles Data Warehousing Concepts

© The Data Warehousing Institute 1-1

Module 1

Data Warehousing Concepts

Topic Page Data Warehousing Basics 1-2

The Data Warehousing Application 1-10

Warehousing Data Stores 1-16

The Data Warehousing Process 1-30

Data Warehousing Deliverables 1-34

The Data Warehousing Program 1-36

Readiness Assessment 1-38

Data Warehousing Concepts TDWI Data Warehousing Concepts and Principles

1-2 © The Data Warehousing Institute

Data Warehousing Basics Understanding Data, Information, and Knowledge

Outcomeachievement, discovery

Actioninsight, resolve, decision, innovation

Knowledgerecall, experience, instinct, beliefs

Datadescriptive, quantitative, qualitative

Informationfacts, metrics

impact

done by software& databases

done by people

realizes business valueOutcome

achievement, discovery





impact



done by peopledone by people

realizes business valuerealizes business value



Data Warehousing Basics Understanding Data, Information, and Knowledge

DATA Data is composed of individual and discrete facts that collect descriptive, quantitative, and qualitative values of business interest. Data warehousing involves two types of data – operational data which describe the day-to-day events and transactions of the business, and informational data that are reconciled, integrated, and cleansed to constitute the raw material from which information is constructed.

INFORMATION Information is an organized collection of data presented in a specific and

meaningful context. The purpose of business information is to inform people and processes – to provide facts and metrics vital to the processes and useful to the people who carry out those processes. Information adds to the collection of knowledge that is available to business people and business processes.

KNOWLEDGE Knowledge is a personal and individual thing. Here we leave the realm of

what computers and software do, and enter the domain of what people do. Knowledge encompasses the familiarity, awareness, understanding, and perceptions of a person about a given subject. Knowledge is gained through many channels including study, recall, experience, instinct, and beliefs. These factors are different for each person, thus the knowledge of every individual is unique

ACTIONS AND OUTCOMES

Action is a process of doing something. Effective action is the process of doing the right thing. It is described as a process because we need to look beyond the event of doing and consider the activities and behaviors that lead to that event. Any combination of insight, resolve, decision, and innovation may drive a person to act – the “doing” part of action. Outcomes are the results of actions. Favorable business outcomes are generally those that reduce cost, save time, optimize resources, increase revenue, satisfy customers, or otherwise help to fulfill the business mission and goals.

IMPACT AND VALUE

Value is realized at the bottom line of the business – when outcomes reduce cost or increase revenue either directly or indirectly. The value of an action is determined by the outcomes produced. The value of information is derived through contribution to valued action – providing support for insight, resolve, decision, and innovation. The value of the data warehouse depends entirely on the value of the information services that it delivers.



Warehousing Data Stores Data Store Responsibilities

Data Marts(cubes, views, web reports, spreadsheets, etc.)

SourceData

source to warehouseETLs

queries & analysis

Business Intelligence Tools

DataStaging

DataWarehouse

Data Marts(star-schema, cubes, views, web reports, spreadsheets, etc.)

intake

integration

distribution

delivery

access Data Marts(cubes, views, web reports, spreadsheets, etc.)

SourceData


queries & analysis


DataStaging

DataWarehouse


Data Marts(cubes, views, web reports, spreadsheets, etc.)

SourceData


queries & analysis


DataStaging

DataWarehouse

SourceData


queries & analysis


DataStaging

DataWarehouse

DataStaging

DataStaging

DataWarehouse

DataWarehouse


intake

integration

distribution

delivery

access

intake

integration

distribution

delivery

access



Warehousing Data Stores Data Store Responsibilities

THE ROLES Every data warehousing environment, regardless of architecture and flow of data, must provide for five roles to be complete. Different architectures assign these roles to data stores in various ways.

INTAKE Data stores with intake responsibility receive data into warehousing

environment. Data is acquired from multiple source systems, of varying technologies, at different frequencies, and into numerous warehousing files and/or tables. Further, the data typically requires many and diverse transformations. Most data is extracted from operational systems whose data is most certainly not all clean, error-free and complete. Data cleansing is commonly performed as part of the intake process to ensure completeness and correctness of data.

INTEGRATION Integration describes how the data fits together. The challenge for the

warehousing architect is to design and implement consistent and interconnected data that provides readily accessible, meaningful business information. Integration occurs at many levels – “the key level, the attribute level, the definition level, the structural level, and so forth …” (Data Warehouse Types, www.billinmon.com) Additional data cleansing processes, beyond those performed at intake, may be required to achieve desired levels of data integration.

DISTRIBUTION Data stores with distribution responsibility serve as long-term information

assets with broad scope. Distribution is the progression of consistent data from such a data store to those data stores designed to address specific business needs for decision support and analysis.

DELIVERY Data stores with delivery responsibility combine data as “in business

context” information structures to present to business units who need it. Delivery is facilitated by a host of technologies and related tools - data marts, data views, multidimensional cubes, web reports, spreadsheets, queries, etc.

ACCESS Data stores with access responsibility are those that provide business

retrieval of integrated data – typically the targets of a distribution process. Access-optimized data stores are biased toward easy of understanding and navigation by business users.



Warehousing Data Stores The Data Warehouse

SourceData


queries & analysis


The “Kimball”Data Warehouse

intake

integration

distribution

delivery

access

Data Warehouse

Data Marts(star-schema and/or cubes)

The “Inmon”Data Warehouse

SourceData


SourceData


queries & analysis


DataWarehouse

intake

integration

distribution

SourceData


queries & analysis



intake

integration

distribution

delivery

access

Data Warehouse


SourceData


SourceData


queries & analysis



intake

integration

distribution

delivery

access

intake

integration

distribution

delivery

access

Data WarehouseData Warehouse



SourceData


SourceData


queries & analysis


DataWarehouse

intake

integration

distribution


SourceData


SourceData


SourceData


queries & analysis


DataWarehouse

intake

integration

distribution

intake

integration

distribution



Warehousing Data Stores The Data Warehouse

CENTRAL DATA WAREHOUSE (HUB)

As previously discussed, Inmon defines a data warehouse “a subject-oriented, integrated, non-volatile, time-variant, collection of data organized to support management needs.” (W. H. Inmon, Database Newsletter, July/August 1992) The intent of this definition is that the data warehouse serves as a single-source hub of integrated data upon which all downstream data stores are dependent. The Inmon data warehouse has roles of intake, integration, and distribution.

KIMBALL’S DEFINITION (BUS)

Kimball defines the warehouse as “nothing more than the union of all the constituent data marts.” (Ralph Kimball, et. al, The Data Warehouse Life Cycle Toolkit, Wiley Computer Publishing, 1998) This definition contradicts the concept of the data warehouse as a single-source hub. The Kimball data warehouse assumes all data store roles -- intake, integration, distribution, access, and delivery

DIFFERENCES IN PRACTICE

Given these two predominant definitions of the data warehouse - Inmon’s (hub-and-spoke architecture) and Kimball’s (bus architecture), what are the implications with regard to the five roles of a data store – intake, integration, distribution, access and delivery?

Inmon Warehouse Kimball Warehouse intake fills the intake role, but may be

downstream from staging area Fills the intake role – downstream from “backroom” transient staging

integration Primary integrated data store with data at the atomic level

Integration through standards and conformity of data marts

distribution Designed and optimized for distribution to data marts

Distribution is insignificant because data marts are a subset of the data warehouse

access May provide limited data access to some “power” users

Specifically designed for business access and analysis

delivery Not designed or intended for delivery

Supports delivery of information to the business



Data Warehousing Deliverables Results of Architecture, Implementation & Operation Activities

data warehousing program charterdata warehousing readiness assessmentdefined business architecturedefined data architecturedefined technology architecturedefined project architecturedefined organizational architecture

Architecture

project planstarget data modelsdata warehousing process modelsdeployed technologywarehousing databasesdata acquisition processesdata transformation processesdata transport & load processespopulated warehousing databasesbusiness analysis applicationsdelivered data warehousing capabilities

Implementation

business servicesdata refreshmanaged platformsmanaged environmentcustomer servicemanaged qualitymanaged infrastructure

Operation


Architecture


Architecture


Implementation


Implementation


Operation


Operation



Data Warehousing Deliverables Results of Architecture, Implementation & Operation Activities

ARCHITECTURE RESULTS

Architectural activities establish the standards, conventions, and guidelines that ensure consistency and integration among results of multiple implementation projects. Architectural work begins by defining a warehousing program and assessing organizational readiness. Architecture is broad in scope and focused on analysis and design in the following areas: • Business Architecture – Understanding of business goals, drivers, and

information needs. • Data Architecture – Understanding of source data. Requirements and

standards for warehousing data and warehouse metadata. • Technology Architecture – Identification of standards for hardware,

software, and communications technology. Specification of the data warehousing toolset.

• Project Architecture – Incremental development plan for the data warehouse. Defined scope of each increment. Sequence and dependencies among increments.

• Organizational Architecture – Identification of training, support, and communications responsibilities.

IMPLEMENTATION RESULTS

Where architecture is broad in scope, implementation narrows the scope to that of a single increment. Each increment is defined as a project that focuses on design, construction, and deployment of warehousing products including: • Warehousing Databases – Data models and implemented databases

for staging data, data warehouse, and data marts. • Warehousing Processes – Source –to-target mapping, specification of

data transformation rules, and development of processes to move data through the warehousing environment.

• Business Analysis Applications – Standard queries, decision support systems (DSS), warehouse published reports, and other standard means of receiving information from the data warehouse.

OPERATION RESULTS

Operation is the phase where data warehousing delivers value. That value is realized through business services that provide data and information and enable confident decisions and positive actions. Training, support, and administration are also key elements of data warehouse operation.

TDWI Data Warehousing Concepts and Principles Data Warehousing Architecture


Module 2

Data Warehousing Architecture

Topic Page Business Architecture 2-2

Data Architecture 2-10

Technology Architecture 2-46

Project Architecture 2-48

Organizational Architecture 2-58

This page intentionally left blank.

Data Warehousing Architecture TDWI Data Warehousing Concepts and Principles


Business Architecture Business Processes

activities

inputs

product customerssourcesworkforce

events / transactions

business

process

• which processes are in scope of the warehousing program?• who (customer, source, workforce) needs information?• which business process components are information subjects?

• how can inputs be optimized?• how can activities be streamlined?• who can the workforce contribute?• how can suppliers contribute?• how can events be managed?• how can product value be enhanced?

activities

inputs

product customerssourcesworkforce


business

processactivities

inputs

product customerssources inputs

inputs

product customerscustomerssourcessourcesworkforce


business

process

• which processes are in scope of the warehousing program?• who (customer, source, workforce) needs information?• which business process components are information subjects?

• how can inputs be optimized?• how can activities be streamlined?• who can the workforce contribute?• how can suppliers contribute?• how can events be managed?• how can product value be enhanced?



Business Architecture Business Processes

UNDERSTANDING BUSINESS PROCESSES

Business processes are the things that a business does to produce its products, deliver its services, manage its infrastructure, etc. Every business process can be understood in terms of the components of that process: • the product that the process produces, • the customer who uses the product, • the inputs that are needed to produce the product, • the sources/suppliers that provide the inputs, • the activities that comprise the process, • the actors who perform the activities, • the events that drive the activities. Recognizing which processes will be information-enabled through data warehousing, and which process components will become subjects of warehousing data, offers valuable input to all phases of data warehouse planning, development, and operation.



Data Architecture Data Modeling Concepts

• Warehousing Subjects• Business Questions• Facts & Qualifiers• Target Configuration

• Staging, Warehouse, & MartER Models

• Data Mart DDMs

• Staging Area Structure• Warehouse Structure• Relational Mart Structures• Dimensional Mart Structures

• Staging Physical Design• Warehouse Physical Design• Data Mart Physical Designs

(relational & dimensional)

• Implemented WarehousingDatabases

• Source Composition• Source Subjects

• Integrated Source DataModel (ERM)

• Source Data Structure Model

• Source Data Files

• Source Data FileDescriptions

• Business Goals & Drivers• Information Needs

Triage

ContextualModels

ConceptualModels

LogicalModels

StructuralModels

PhysicalModels

ImplementedData



Data Architecture Data Modeling Concepts

FAMILIAR DATA MODELING PRINCIPLES

Like application data modeling, warehouse modeling works well when practiced at multiple levels of abstraction. Modeling either application or warehouse data may develop any or all of: • Contextual Models describing the scope of requirements, establishing

a context for analysis. • Conceptual Models describing requirements without consideration for

computer implementation. • Logical Models describing data from a computer system perspective,

yet free of any implementation platform specifics. • Structural Models specifying data structures that account for variables

of access, navigation, security, distribution, and time-variance. • Physical Models providing detailed design and specification of data

structures to be implemented using a particular technology. WAREHOUSE MODELING DIFFERENCES

Even the most experienced application data modelers are challenged by early warehouse modeling experiences. New issues, terminology, and techniques combine to make warehouse data modeling more complex than application data modeling. The primary differences include:

This Facet of Warehouse Modeling …

Differs from Application Modeling in This Way …

Multiple Data Types

Both source data and warehousing data need to be modeled. Each is modeled separately, and they are associated through a technique called “triage.”

Multiple Ways to Use Warehouse Data

Warehouse data uses range from publishing and managed query to complex OLAP applications and data mining. The ideal data structure depends on planned uses of the data.

Multiple Ways to Organize the Data

Warehouse databases may be organized relationally, dimensionally, or with a combination of the two techniques. The ideal organization depends on both the planned uses of the data and the characteristics of the data.

Multiple Modeling Techniques The complexities of warehouse data modeling require that many modeling techniques be used. Matrix models, E/R models, subject models, dimensional models, star-schema, and snowflake-schema are all used to meet various data modeling needs.

Planned and Managed Redundancy

Redundancy, typically avoided in application databases, is an asset to warehouse databases. Planning and managing redundancy is a key skill for warehouse data modelers.

Large Data Volumes Redundancy and time-variance combine to make a very large database (VLDB) a common warehouse consideration. Optimizing for data volumes and database size is a common requirement of warehouse modeling.



Data Architecture Integration and Data Flow Standards

DataSources

IntegrationHub

DataMart Data

Mart

DataMart

Hub and Spoke Integration

DataSources

IntegrationHub

DataMart Data

Mart

DataMart

Hub and Spoke Integration

DataSources

Bus Integration

Integration Bus

DataMart Data

Mart

DataMart

DataSources

DataSources

Bus Integration

Integration BusIntegration Bus

DataMart Data

Mart

DataMart

DataMartDataMart Data

MartDataMart

DataMartDataMart



Data Architecture Integration and Data Flow Standards

HUB-AND-SPOKE INTEGRATION

The hub-and-spoke architecture provides a single integrated and consistent source of data from which data marts are populated. The warehouse structure is defined through enterprise modeling (top down methodology). The ETL processes acquire the data from the sources, transform the data in accordance with established enterprise-wide business rules, and load the hub data store (central data warehouse or persistent staging area). The strength of this architecture is enforced integration of data.

BUS INTEGRATION The Bus Architecture relies on the development of conformed data marts

populated directly from the operational sources or through a transient staging area. Data consistency from source-to-mart and mart-to-mart are achieved through applying conventions and standards (conformed facts and dimensions) as the data marts are populated. The strength of this architecture is consistency without the overhead of the central data warehouse.



Project Architecture Methodology

Enterprise Modeling & Architecture

Incremental Development Planning

Data Warehouse Design & Development

Data Mart Design & Development

Operation & Support

Incremental Deployment

Top-Down Development

Data Mart Deployment


Operation & Support

Identify Business Area Scope

Bottom-Up Development

Incremental Enterprise Modeling

Integration StructureDesign & Development

IncrementalDeployment

Data Warehouse / MartDesign & Development

Operation & Support

Incremental DevelopmentPlanning

Identify BusinessArea Scope

Hybrid Methods





Operation & Support






Operation & Support


Top-Down Development



Operation & Support




Operation & Support


Bottom-Up Development





Operation & Support







Operation & Support







Operation & Support



Hybrid Methods



Project Architecture Methodology

TOP-DOWN Top-down approaches are also commonly called enterprise approaches. Top-down data warehouse development begins at the enterprise, and typically emphasizes the data warehouse as a primary integrated information resource. Data warehouse structure is determined through enterprise modeling. Content is determined by a combination of business information needs and available source data. Top-down approaches are generally associated with longer start-up times due to the need for enterprise perspective.

BOTTOM-UP Bottom-up approaches begin with business information needs for a single

business unit or limited business domain. Bottom-up methods are most compatible with bus integration approaches, using conformity instead of an enterprise repository to achieve integration. Bottom-up development generally trades strength of an integration hub for the benefits of quick start-up and rapid deployment.

BALANCING ENTERPRISE & BUSINESS UNIT FOCUS

Hybrid approaches combine some elements of bottom-up development with some from top-down methods. The objective of a hybrid approach is rapid development within an enterprise context. A typical hybrid approach quickly develops a skeletal enterprise model before beginning iterative development of data marts. The data warehouse is populated only as data is needed by data marts, and is sometimes constructed in a retrofit mode after data marts have been deployed. Metadata consistency and conformed dimensions are the initial integration tools, with the data warehouse being a secondary means of integration



Organizational Architecture Program, Project & Operations Roles

sponsorship program management data governance

metadata management architecture specification quality management

busin

ess r

ules s

pecif

icatio

nbusiness requirements definition

integration design database development ETL development

data mart development BI application development

proje

ct ma

nage

ment

source data analysis

data integration & cleansingdata access, analysis, & mining

business metrics usagesystem & database administrationprocess execution & monitoring

training & support

BI Program

BI Projects

BI Operations

sponsorship program management data governance

metadata management architecture specification quality management

busin

ess r

ules s

pecif

icatio

nbusiness requirements definition

integration design database development ETL development

data mart development BI application development

proje

ct ma

nage

ment

source data analysis

data integration & cleansingdata access, analysis, & mining

business metrics usagesystem & database administrationprocess execution & monitoring

training & support

BI Program

BI Projects

BI Operations



Organizational Architecture Program, Project & Operations Roles

ROLES AND RESPONSIBILITIES

The program, project, and operation activities of data warehousing are different from those of developing and supporting operational systems. The work is different; therefore the roles and responsibilities are different. Data warehousing has different goals and challenges. It demands different kinds of organizations and teams. Common data warehousing roles and responsibilities include:

BI Program Roles & Responsibilities

Program Management Managing business/IT relationship, multiple dependent projects, issue resolution, etc. Sponsorship Advocacy, political will, resource acquisition, issue resolution, expectation setting, etc. Data Governance Data definitions, business rules alignment, data quality management, access authorization, etc. Business Rules Specification Business basis for data rules about content, relationships, correctness, integrity, etc. Business Requirements Definition Requirements for data & information, service levels, quality & reliability, etc. Architecture Specification Frameworks & standards for business alignment, data, technology, projects, etc. Quality Management Beyond data quality – quality of information, delivery, interface, reporting, services, etc. Meta Data Management Meta data strategy, meta data implementation, meta data content, etc.

BI Project Roles & Responsibilities Project Management Work breakdown, scheduling, resource allocation, deliverables, deployment, etc. Integration Design Data source selection, source/target mapping, transformation rules, populating databases Database Development Logical and physical database design, database specification and creation ETL Development Analysis, design, construction, and deployment of data movement processes Source Data Analysis Data profiling, source content analysis, source data modeling Data Mart Development Analysis, design, construction, and deployment of data marts BI Application Development Analysis, design, construction, and deployment of information services & analytic applications

BI Operations Roles Data Integration & Cleansing Maintenance and support of data migration processes; Continuous data quality management Data Access, Analysis, & Mining Access and application of data to make business decisions Business Metrics Usage Application of business measures to drive business actions System & Database Administration Installation, configuration, and management of BI operating platforms Process Execution & Monitoring Scheduling, execution, verification, and support of data warehousing processes Training & Support Customer care activities for BI customers

TDWI Data Warehousing Concepts and Principles Data Warehouse Implementation

© Th Data Warehousing Institute 3-1

Module 3 Data Warehouse Implementation

Topic Page Implementation Planning 3-2

Warehouse Data Modeling 3-8

The Warehouse Process Model 3-22

Deployed Technology 3-40

Implementation Components 3-44

Delivery Results 3-48


Data Warehouse Implementation TDWI Data Warehousing Concepts and Principles


Warehouse Data Modeling Logical Models of Dimensional Data

• Warehousing Subjects• Business Questions• Facts & Qualifiers• Target Configuration

• Staging, Warehouse, & MartER Models

• Data Mart DDMs

• Staging Area Structure• Warehouse Structure• Relational Mart Structures• Dimensional Mart Structures

• Source Composition• Source Subjects

• Integrated Source Data Model(ERM)

• Source Data Structure Model

• Business Goals & Drivers• Information Needs

Triage

ContextualModels

ConceptualModels

LogicalModels

StructuralModels

• Staging Physical Design• Warehouse Physical Design• Data Mart Physical Designs

(relational & dimensional)

• Source Data FileDescriptions

customer-count

household-count

SIZE OF

CUSTOMER BASE

ProductLOB

lob-codelob-name

PRODUCT LINE

line-code

line-description

PRODUCT

product-idproduct-descproduct-name

Time

MONTH

month-number

QUARTER

quarter-number

YEARyear-number

Geographic Area

REGIONrgn-codergn-name

DISTRICT

dist-number

dist-name

ZONEzone-number

zone-name



Warehouse Data Modeling Logical Models of Dimensional Data

EXAMPLE The diagram on the facing page illustrates an example of a dimensional data model at the logical level. This example shows a data mart whose purpose is to measure the size of the customer base.



The Warehouse Process Model Source/Target Maps

member-numbermembership-typedate-joineddate-last-renewedterm-last-reneweddate-of-last-activitylast-namefirst-namebusiness-nameaddresscity-and-statezip-code

MEMB

ERSH

IP M

ASTE

R

date-timeterminal-idtransaction-idline-numberSKU

Files

/Tab

les an

d Fi

elds f

rom

Sou

rce S

truct

ural

Mode

l

Tables and Data Elements from Target Structural Model

trans

actio

n-da

tetra

nsac

tion-

time

stor

e-nu

mbe

rtra

nsac

tion-

amt

regi

ster

-idtra

nsac

tion-

stat

uspa

ymen

t-met

hod

prod

uct-c

ode

prod

uct-S

KUpr

oduc

t-typ

epr

oduc

t-des

crip

.

mem

ber n

umbe

rcu

stom

er n

ame

mem

bers

hip

date

rene

wal d

ate

cust

omer

addr

ess

SALES TRANSACTIONCUSTOMER PRODUCT

POIN

T-OF

-SAL

E DE

TAIL



The Warehouse Process Model Source/Target Maps

SOURCE AND TARGET DATA ASSOCIATIONS

Source/target mapping develops detailed understanding of the associations between source data and target data. Mapping may occur at three levels: • Mapping entities to understand the business associations • Mapping tables and files to understand associations among data

stores • Mapping columns and fields to understand associations at the data

element level The focus of this mapping is on what associations exist, without examining which are the most desirable sources or how the data might be translated.



The Warehouse Process Model Data Transformation Rules

member-numbermembership-typedate-joineddate-last-renewedterm-last-reneweddate-of-last-activitylast-namefirst-namebusiness-nameaddresscity-and-statezip-code

MEMB

ERSH

IP M

ASTE

R

date-timeterminal-idtransaction-idline-numberSKU

Files

/Tab

les an

d Fi

elds f

rom

Sou

rce S

truct

ural

Mode

l

Tables and Data Elements from Target Structural Model

trans

actio

n-da

tetra

nsac

tion-

time

stor

e-nu

mbe

rtra

nsac

tion-

amt

regi

ster

-idtra

nsac

tion-

stat

uspa

ymen

t-met

hod

prod

uct-c

ode

prod

uct-S

KUpr

oduc

t-typ

epr

oduc

t-des

crip

.

mem

ber n

umbe

rcu

stom

er n

ame

mem

bers

hip

date

rene

wal d

ate

cust

omer

addr

ess

SALES TRANSACTIONCUSTOMER PRODUCT

POIN

T-OF

-SAL

E DE

TAIL

DTR027 (Default Membership Type) If membership-type is null or invalid assume “family” membership

DTR009 (Translate StatusDTR008 (Derive Name) If membership-type is business use business-name else concatenate last-name and first-name separated by a commaDTR009 (Translate Status

cells expand to identify transformationsby type & name

logic oftransformationsis separatelydocumented

Cleansing DTR027 (Default Value)

Derivation DTR008 (Derive Name)



The Warehouse Process Model Data Transformation Rules

DETAILED SPECIFICATION

Specification of data transformations develops a large set of details about how source data is to be processed prior to loading of a warehousing database. Documenting data transformation must address both the identification of what transformations are needed, and the logic of the transformation process. Documenting which transformations occur can readily be achieved by extending the source/target maps. View the set of logic for each transformation as a unique rule, and develop a convention for naming these rules. As each transformation need is identified, assign a name and place that name in the appropriate cell of a source/target map. Then document the logic of each transformation rule. For each source/target association consider possible rules for each of the transformation types. In addition, consider need for data cleansing. Although data clean-up is not a unique transformation rule type, it is a common reason for filtering, conversion, and derivation.

Transform Need Description

Specify Selection Requirements

Identifies and describes the selection processes needed to choose among multiple sources. The objective is to select the best data to be used for warehouse population. Selection requirements may exist at both data store and date element levels.

Specify Filtering Requirements

Identifies and describes the filtering processes needed to choose records from a source file (or rows from a source table) to be used for data warehouse population.

Specify Conversion and Translation Requirements

Identifies and describes the conversion and translation processes which change the formats and values of data elements. Conversion processing achieves consistency of formats and value sets among data extracted from multiple sources. Translation processes change data formats and values from encoded and cryptic to descriptive and meaningful.

Specify Derivation and Summarization Requirements

Identifies and describes needed derivation processes used to develop a value for a single data element by applying logic to the values of some other data elements. It also identifies and describes the processes through which summary data values are created.

Specify Clean-up Requirements

Identifies and describes the clean-up processing needed to ensure quality and integrity of the data that is placed into the data warehouse. Clean-up needs may exist at both data record and data element levels. Among the issues of clean-up processing are intra- and inter-record consistency checking, and decisions regarding elements with null values or invalid values.



Deployed Technology Range and Roles of Technology

web desktop email wireless voice

B2E

Porta

l (in

trane

t)

B2B & B2C Portals (internet/extranet)

Analytic ApplicationsBPM (scorecards & dashboards)

CRM Analytics Supply Chain Analytics Operations Analytics

Analytic Apps DevelopmentTools, Packages, Templates Collaboration

E-mail, Groupware, Workflow

Data Access & AnalysisQuery, Reporting, OLAP, Mining, Forecasting Text Analysis

Text Search & Text Mining

Data Warehouse / Data Marts Content Management

Data IntegrationModeling, Mapping, Cleansing, ETL

Data ResourcesOperational Systems, Documents, Images, External Data, Audio/Visual

Infra

stru

ctur

eSt

orag

e, Se

rvers,

Data

base

s, Me

tadata

, Adm

inistr

ation

& M

anag

emen

t, Netw

orkin

g

web desktop email wireless voice

B2E

Porta

l (in

trane

t)

B2B & B2C Portals (internet/extranet)

Analytic ApplicationsBPM (scorecards & dashboards)

CRM Analytics Supply Chain Analytics Operations Analytics

Analytic Apps DevelopmentTools, Packages, Templates Collaboration

E-mail, Groupware, Workflow

Data Access & AnalysisQuery, Reporting, OLAP, Mining, Forecasting Text Analysis

Text Search & Text Mining

Data Warehouse / Data Marts Content Management

Data IntegrationModeling, Mapping, Cleansing, ETL

Data ResourcesOperational Systems, Documents, Images, External Data, Audio/Visual

Infra

stru

ctur

eSt

orag

e, Se

rvers,

Data

base

s, Me

tadata

, Adm

inistr

ation

& M

anag

emen

t, Netw

orkin

g



Deployed Technology Range and Roles of Technology

TECHNOLOGY ROLES AND RELATIONSHIPS

The technology framework illustrates classes of tools and technology from infrastructure through information delivery. This framework includes established and mainstream technologies as well as emerging technologies (content management, text analysis, text mining, collaboration, etc.) that are gaining significance in data warehousing. The major technology classes are:

Delivery Delivery media includes web portals, desktop clients, email, wireless, voice print, pager and fax.

Delivery technology sets include (1) B2E Portal – intranet business-to-enterprise delivery to the workforce, (2) B2B Portal– internet business-to-delivery to vendors, customers, partners and anyone with internet access, (3) B2C Portal – extranet business-to-customer delivery.

Analytic Applications Analytic applications are the technology components of business applications, ranging from static reporting to dashboards and scorecards. They place information into business function context, i.e. Customer Relationship Management (CRM), Supply Chain Management (SCM), Business Performance Management (BPM), etc.

Analytic Application Development Tools

Tools, templates, and packaged applications to quickly build views, reports, dashboards, scorecards, and other applications to deliver information in context of a business function or business process.

Collaboration Web applications to support employees, partners, customers, vendors and others to collaborate on documents, share business metrics, manage content, and work collectively. While reporting is still dominant today, collaboration capabilities will grow as the technology and market place mature.

Data Access & Analysis Data access and analysis tools are today’s most common delivery technologies. Unlike analytic applications, these tools focus on data before information, and they provide less business context than analytic applications. The most widely-used tools include managed reporting, query, and OLAP.

Text Analysis Text analysis tools use semantics and statistical techniques to identify, tag, and select relevant content from text documents. Parsing, pattern recognition, natural language processing and other advanced techniques are used to transform unstructured text into data and/or information structures.

Data Warehouse / Marts Data warehouses and data marts integrate and reconcile data from multiple data sources. Their purpose is to prepare data to serve as the raw material from which information is created. Regardless of the multiple definitions of data warehouse and data mart that are used in the industry, all warehouses and marts exist primarily to serve this purpose.

Content Management Content management technology first emerged as an internet technology – to support management of content-rich web sites. Uses of the technology in BI are emerging as the industry evolves from data warehousing to business intelligence, and from integration of structured data integrating all types of business information resources. Basic content management functions include indexing, searching, and retrieval.

Data Resources This class includes all sources from which data can be acquired. When both internal and external data are considered, and when both structured and unstructured data are included, the range of possible source technologies becomes exceptionally broad.

Infrastructure This technology class describes the underlying hardware, software, networking, administration and support structures upon which systems and data sources are constructed and operated.



Delivery Results Data and Information Services

The right kinds of services matched to the customer’s roles, responsibilities, and experience level

ExpertOccasional

Use Beginner

Data Access andInformation DeliveryServices

Analysis & ReportingServices

Training Services

Support Services

RegularUse

Business ManagerKnowledge Worker

Executive

array of

BI service

s

ExpertOccasional

Use Beginner

Data Access andInformation DeliveryServices

Analysis & ReportingServices

Training Services

Support Services

RegularUse

Business ManagerKnowledge Worker

Executive

array of

BI service

sarray of

BI service

s



Delivery Results Data and Information Services

MEETING CUSTOMER NEEDS

A mature data warehousing environment includes a robust set of services that support the goal of delivering the right services to the right people at the right time. A three-dimensional view of the services array is useful to classify services and to assess customer needs and match them with available services. The services dimensions are: • Classification of customers as

o Knowledge workers who carry out the day-to-day activities of the business

o Managers responsible for performance of individual business processes

o Executives responsible for business performance across many business processes

• Classification of customer experience as

o Experts who use the data warehouse regularly and have a high level of computer and analytic skills combined with an intimate knowledge of data warehouse content

o Regular users of the data warehouse with moderate computer and analytic skills combined with a working knowledge of data warehouse content

o Occasional users of the data warehouse who may have necessary computer and analytic skills, but have limited knowledge of data warehouse content

o Beginners with little or no knowledge of data warehouse content, and who may have limited computer or analytic skills

• Classification of services as

o Data access and information delivery services that make data and information available to the business.

o Analysis and reporting services that deliver analytic applications of greater complexity than simple data access and information delivery.

o Training services that develop customer skill and ability to use the data warehouse, with a goal of making each customer self-sufficient.

o Support services that augment the services culture, enhance communications with customers, and ensure rapid resolution of problems.

TDWI Data Warehousing Concepts and Principles Data Warehouse Operation


Module 4

Data Warehouse Operation

Topic Page Business Services 4-2

Data Warehouse Administration 4-6

Managed Quality 4-14

Managed Infrastructure 4-16


Data Warehouse Operation TDWI Data Warehousing Concepts and Principles


Business Services Valuable and Sustainable Services

VALUABLE AND SUSTAINABLE

DATA & INFORMATION SERVICES

FOR THE BUSINESSbusiness servicesdata refreshmanaged platformsmanaged environmentcustomer servicemanaged qualitymanaged infrastructure

Operation






impact

VALUABLE AND SUSTAINABLE

DATA & INFORMATION SERVICES

FOR THE BUSINESSbusiness servicesdata refreshmanaged platformsmanaged environmentcustomer servicemanaged qualitymanaged infrastructure

Operation


Operation






impact






impact



Business Services Valuable and Sustainable Services

WAREHOUSING FOR THE LONG TERM

Sustaining the data warehouse demands a commitment to delivering reliable and valuable business services in an environment of high-frequency change. Value is sustained by ensuring continuous alignment with changing business needs and with a changing customer base. Reliability is sustained by attention to all of the “under the hood” components upon which the services depend including: • Regular, routine, and dependable data refresh despite changing data

sources and systems. • Effectively managed technology platforms from data acquisition to

information delivery in a climate of rapid technological change. • Managed environment including security, growth, capacity planning,

and configuration management. • Customer service including support, help desk, and training services. • Continuous quality management for all aspects of quality – business

quality, data and information quality, and technical quality. • Actively managed infrastructure that ensures continued alignment of

people, processes, and technology for optimum business value.



Managed Quality Dimensions of Quality

what

leve

l of n

eeds

and

expe

ctat

ions

?

what

curre

nt le

vel o

f ser

vice?

focus on business driversalignment with business strategies

enabling of business tactics

understanding of purpose, content, & servicesaccess to needed business information

satisfaction with information availability and reliability

reach into the business communityrange of data and services

maneuverability as change occurscapability to use, adapt, extend & evolve business intelligence

what

leve

l of n

eeds

and

expe

ctat

ions

?wh

at le

vel o

f nee

ds an

d ex

pect

atio

ns?

what

curre

nt le

vel o

f ser

vice?

what

curre

nt le

vel o

f ser

vice?















Managed Quality Dimensions of Quality

QUALITY IMPROVEMENT

Quality, as with any other aspect of business, is effectively managed with measures and metrics. A metrics foundation for quality management includes both measures of product quality and measures of the process that produces the product. In the case of business intelligence, the products are BI results – information delivered to the business, analytics used by the business, actions and outcomes enabled through BI, etc. The processes are those necessary to execute the entire chain of events from data warehousing to business action, and to sustain a BI program over time. Product measures are used to detect defects in BI products and to improve those products. Process measures help to identify causes of defects and prevent reoccurrence through process improvement. A mature quality process regularly adjusts quality targets to achieve continuous improvement.

DIMENSIONS OF QUALITY

Business intelligence quality is much more than simple data quality. Data quality is, in fact, a relatively small and easy piece of the overall quality domain. BI quality is measured and managed in three major categories: • Business Quality directly affects the business value derived from BI,

and the economic success of the BI program. • Information Quality is related to acceptance and use of BI products

– the extent to which BI customers value those products. Information quality is a significant factor in political success of BI.

• Technical Quality involves choosing the right technologies,

configuring multiple technologies to work well together, and using the right tools for the right job. High-quality implementation of technology is typically unnoticed by the business. Low-quality, however, is highly visible and directly affects overall acceptance, usage, trust, value realization, and sustainability of a BI program.



Managed Infrastructure Processes, Technology, and People

program managementchange managementquality management

data governancedevelopment methodologies

project managementdata warehouse administration

metadata managementdata warehousing tools & technology

BI tools & technologyinfrastructure tools & technology

BI roles & responsibilitiesBI organizations

program managementchange managementquality management

data governancedevelopment methodologies

project managementdata warehouse administration

metadata managementdata warehousing tools & technology

BI tools & technologyinfrastructure tools & technology

BI roles & responsibilitiesBI organizations



Managed Infrastructure Processes, Technology, and People

COMPLEX INFRASTRUCTURE

Infrastructure is the foundation upon which BI operates and grows. While the infrastructure supports development, it’s more critical role is in operating and sustaining BI solutions. Operation and sustenance are both more demanding and of longer duration than development. An effective BI infrastructure is one in which processes, technology, and people work seamlessly to support a BI culture and to realize business value from BI solutions.

PROCESS This course has already discussed the analytics processes of BI. When

successful, BI becomes a key component in decision making processes. It depends, however, on many other processes to achieve this level of success. The process components of BI infrastructure are program management, change management, data governance, development methodology, project management, data warehouse administration, and metadata management.

TECHNOLOGY While technology can’t create BI, neither can BI be created without use of

technology. Blending the right technologies with the process and people components of BI is a key to success. Technology infrastructure includes data warehousing tools, BI tools, and enabling/infrastructure hardware and software.

PEOPLE People are integral to effective BI. Neither processes nor technology can

deliver value independently of the knowledge, decisions, and actions of people. Human infrastructure is arguably the single most important of all BI infrastructure categories. Identifying the right set of roles and responsibilities, assigning them to people with the right skills, and constructing the right kinds of organizations and relationships are all critical to BI success.

TDWI Data Warehousing Concepts and Principles Summary and Conclusions


Module 5

Summary and Conclusions

Topic Page Common Mistakes 5-2

References and Resources 5-6

TDWI Data Warehousing Concepts and Principles Summary and Conclusions


Common Mistakes From TDWI’s 10 Mistakes Series

An effective project manager will not … 1 Accept an unrealistic schedule. 2 Take on a failing project. 3 Launch a project with a dysfunctional team. 4 Choose the wrong sponsor. 5 Accept unrealistic expectations. 6 Expand the project scope. 7 Skip the project plan. 8 Fail to put the project agreement in writing. 9 Let IT drive the project.

10 Give others authority to select software. 11 Market the project alone.

Effective team-builders will avoid … 1 Hiring yourself. 2 Squelching disagreement. 3 Confusing titles with roles and responsibilities. 4 “Talking the walk.” 5 Thinking one size fits all. 6 Pointing fingers. 7 Interviewing only for technical skills. 8 Limiting leadership. 9 Becoming too task focused.

10 Believing that all decisions are created equal.

An effective data modeler will avoid … 1 Not gathering business requirements. 2 Saving time by not creating a subject area model. 3 Delivering normalized tables to drive data mart design. 4 Designing the staging process for ease of developers at end-user expense. 5 De-normalizing without starting from a fully normalized data model. 6 Allowing users to drive the level of detail. 7 Not modeling all levels of a multi-tiered warehousing environment. 8 Developing a data model from a list of required data elements. 9 Believing you must choose between relational and dimensional models.

10 Jumping straight into data mart design.

Summary and Conclusions TDWI Data Warehousing Concepts and Principles


References and Resources Publications

BEST BOOKS: INTERNET SITES:

Marco – Building & Managing the Metadata Repository2000, John Wiley & Sons

Moss & Atre - Business Intelligence Roadmap2003, Addison-Wesley

Inmon – Building the Data Warehouse (3rd Edition)2002, John Wiley & Sons

Inmon, Imhoff & Sousa – Corporate Information Factory (2nd Edition)2000, Johy Wiley & Sons

Kimball - The Data Warehouse Toolkit1996, John Wiley & Sons

Kimball, Reeves, Ross & Thornthwaite – The Data Warehouse Lifecycle Toolkit1998, John Wiley & Sons

Adelman & Moss – Data Warehouse Project Management2000, Pearson Education

The Data Warehousing Institute (www.dw-institute.com)Business Intelligence and Data Warehousing

The Data Administration Newsletter (www.tdan.com)Information and Data Management

The Data Warehousing Information Center (www.dwinfocenter.org)Data Warehousing Resources

Inmon Associates, Inc.(www.billinmon.com)The Inmon Approach to Data Warehousing

The Ralph Kimball Group (www.rkimball.com)The Kimball Approach to Data Warehousing