Sie sind auf Seite 1von 17

Business Intelligence:

What Does it Mean to Me?

Tom Peters
B I Practice Lead
. .

Z Y Solutions Corporation

© 2007 – Z Y Solutions Corporation

What is Business Intelligence

• The Challenge
– Getting to the required information quickly and accurately
to make decisions
• The Problem
– Too much information that is not in a consistent usable
format readily available to those who need it
• The Solution
– Business Intelligence

© 2007 – Z Y Solutions Corporation

Components of Business Intelligence

• Data
– Data store collection by individual applications
• Production systems
– Gathering of all data stores
• Data warehousing
• Presentation
– Scorecards & Dashboards
– Enterprise Reporting
– OLAP Analysis
– Advanced & Predictive Analysis
– Alerts & Proactive Notification

© 2007 – Z Y Solutions Corporation


What is a Data Warehouse

• A generic term for a system of storing, retrieving


and managing large amounts of data
• Software often includes sophisticated compression
and hashing techniques for fast searching and
filtering
• Typically remote database containing recent
snapshots of data
• Planners and researchers can freely use this
information without worrying about slowing down
operation of the production database

© 2007 – Z Y Solutions Corporation

What is a Data Mart

• Type of data warehouse designed mainly to address


a specific function or department’s needs
• Often uses aggregation or summarization of the
data to enhance query performance
• Important, however, to maintain the ability to
access the underlying base data to enable drill-
down analysis as necessary

© 2007 – Z Y Solutions Corporation

Do I Data Warehouse?

• Is the production database designed to be carved up


by the desired dimensions?
– Designed for Online Transaction Processing (OLTP)
– Is database index for all dimensions
– Can I create additional indexes
– What affect will vendor updates have on the solution

© 2007 – Z Y Solutions Corporation


Do I Data Warehouse?

• Can the production database, including hardware,


support the user base and query demands
– Will queries slow my production system
– Will the production system priority prevent BI users from
getting timely data

© 2007 – Z Y Solutions Corporation

Do I Data Warehouse?

• Does the production database contain all the


information I need
– Is there additional historical data

© 2007 – Z Y Solutions Corporation

Do I Data Warehouse?

• Does the production database have the ability to


derive the analytics I need
– Does the data need to be enriched
– Is there data to be derived from the data

© 2007 – Z Y Solutions Corporation


Do I Data Warehouse?

• Will I be replacing my production application


– Design around your business NOT your source applications
– Insulate from source applications allow for replacement

© 2007 – Z Y Solutions Corporation

Do I Data Warehouse?

• Do I need to aggregate data due to very high


transactions counts
– Improve performance
– Limit return row count

© 2007 – Z Y Solutions Corporation

Do I Data Warehouse?

• Always a good idea


– Start simple
– Grow as demands increase

© 2007 – Z Y Solutions Corporation


B.I. Terms

• Relational Database (RDB)


– A database that conforms to the relational model
• Relational Database Management System (RDBMS)
– Refers to the software used to create a RDB
• Informix
• Microsoft SQL Server
• Oracle

© 2007 – Z Y Solutions Corporation

B.I. Terms

• Online Transaction Processing (OLTP)


– Operational systems
– High volume data collection

© 2007 – Z Y Solutions Corporation

B.I. Terms

• Online Analytical Processing (OLAP)


– Defined as providing fast access to shared multi-
dimensional data
– Used to generically refer to software and applications that
provide users with the ability to store and access data in
multi-dimensionally

© 2007 – Z Y Solutions Corporation


Types of OLAP

• Multidimensional Online Analytical Processing


(MOLAP)
– MOLAP cubes are built for fast data retrieval and are
optimal for slicing and dicing operations
– Advantages
• Excellent performance
• Can return complex calculations
– Disadvantages
• Limited in scope as definition of cube creates boundaries
• Limited in volume of data

© 2007 – Z Y Solutions Corporation

Types of OLAP

• Relational Online Analytical Processing (ROLAP)


– Manipulates the data stored in the relational database to
give the appearance of MOLAP's slicing and dicing
functionality.
– Advantages
• Does not constrain data
• No data size limitation
• Can leverage functions of RDB
– Disadvantages
• Each request must query the RDB
• ROLAP itself is limited to RDB functionality

© 2007 – Z Y Solutions Corporation

MOLAP vs. ROLAP

MOLAP Analysis ROLAP Analysis

Revenue for Laptop Computers


In 2007
At All Stores
e
Tim

Geography
Products

Revenue for All Electronics


In 2003 and Q1 2004
At Stores in the NE Region

Total Revenue and Costs


In Jan 2004 and Jan 2003
At Top 10 Revenue Stores

MOLAP Manipulations Allow People To Slice- ROLAP Architecture Allows People To ‘Drill Anywhere’ in The
and-Dice a Subset of Data To View It from Many Entire Relational Database – Across All Dimensions and From
Different Perspectives Summaries To Transactional-level Detail

© 2007 – Z Y Solutions Corporation


Types of OLAP

• Hybrid Online Analytical Process (HOLAP)


– HOLAP combines the advantages of both MOLAP and ROLAP
– Advantages
• Combines both MOLAP and ROLAP
– Disadvantages
• Not available from most B.I. vendors or vendor has limited
experience in deploying

© 2007 – Z Y Solutions Corporation

Metadata

• Data about data


• Dimension
– A perspective that can be used to analyze the data
– Dimensions become more useful when there are many descriptive
attributes that can be used for analyzing the data. The term
“Attribute” is often used to describe the extended Dimension
– Examples
• Customer
• Item
• Date
• Fact
– The raw enumerable piece of information about the transaction always
a numeric value (usually aggregatable)
– Examples
• Quantity
• Unit Price
• Count

© 2007 – Z Y Solutions Corporation

Metadata

• Measure
– The product of one or more Facts
– Can be the result of a formula derived from the RDB or the B.I.
Tool analytical engine
– Examples
• Quantity
• Unit Price
• Count
• Quantity * Unit Price
• Average(Unit Price)
• Minimum(Quantity)

© 2007 – Z Y Solutions Corporation


Metadata

• Report Templates
• Filters
• Prompts
• Transformations
• Consolidations
• Drill Maps
• Hierarchy

© 2007 – Z Y Solutions Corporation

Metadata

Location Master

PK Location

Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
PK,FK1 Location

Quantity

Customer Item SKU


Item Master SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Invoice Line
Invoice Header
PK,FK2 Invoice Number
PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Unit Price

© 2007 – Z Y Solutions Corporation

Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
PK,FK1 Location

Quantity

Customer Item SKU


Item Master SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Invoice Line
Invoice Header
PK,FK2 Invoice Number
PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Unit Price

© 2007 – Z Y Solutions Corporation


Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
PK,FK1 Location

Quantity

Customer Item SKU


Item Master SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Invoice Line
Invoice Header
PK,FK2 Invoice Number
PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Unit Price

© 2007 – Z Y Solutions Corporation

Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
Customer PK,FK1 Location

Quantity

Inventory Customer
Item Master Item SKU
Location SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Stock
Keeping
Unit

Invoice Line
Invoice Header
PK,FK2 Invoice Number
PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Unit Price

© 2007 – Z Y Solutions Corporation

Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
Customer PK,FK1 Location

Quantity

Dimensions

Inventory Customer Item SKU


Item Master SKU Master
Location
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Stock
Keeping
Unit

Invoice Line
Invoice Header
PK,FK2 Invoice Number
PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Unit Price

© 2007 – Z Y Solutions Corporation


Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
Customer PK,FK1 Location

Quantity

Dimensions

Inventory Customer
Item Master Item SKU
Location SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Stock
Keeping
Unit

Invoice Line
Inventory Invoice Header
PK,FK2 Invoice Number
Quantity PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Unit Price

Invoice
Quantity

Invoice
Unit Price

© 2007 – Z Y Solutions Corporation

Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
Customer PK,FK1 Location

Quantity

Dimensions

Inventory Customer
Item Master Item SKU
Location SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Stock
Keeping
Unit

Invoice Line
Inventory Invoice Header
PK,FK2 Invoice Number
Quantity PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Facts Unit Price

Invoice
Quantity

Invoice
Unit Price

© 2007 – Z Y Solutions Corporation

Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
Customer PK,FK1 Location

Quantity

Dimensions

Inventory Customer Item SKU


Item Master SKU Master
Location
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Stock
Keeping
Unit

Invoice Line
Inventory Inventory Invoice Header
PK,FK2 Invoice Number
Quantity Quantity PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Facts Unit Price

Invoice Invoice
Quantity Quantity

Invoice Invoice
Unit Price Unit Price

© 2007 – Z Y Solutions Corporation


Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
Customer PK,FK1 Location

Quantity

Dimensions

Inventory Customer
Item Master Item SKU
Location SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Stock
Keeping
Unit

Invoice Line
Inventory Inventory Invoice Header
PK,FK2 Invoice Number
Quantity Quantity PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Facts Unit Price

Invoice Invoice
Quantity Quantity

Invoice
Revenue

Invoice Invoice
Unit Price Unit Price

© 2007 – Z Y Solutions Corporation

Metadata

Location Master

PK Location

Item Order Header Order Line

PK Order Number PK,FK1 Order Number


PK Line Number
Order Date
FK1 Customer Number FK2 Item Number
SKU Location

PK,FK2 SKU
Customer PK,FK1 Location

Quantity

Dimensions

Inventory Customer
Item Master Item SKU
Location SKU Master
PK Customer Number PK,FK1 Item Number
PK Item Number PK SKU
PK,FK2 SKU
Customer Name
Item Description
Customer Address

Stock
Keeping
Unit

Invoice Line
Inventory Inventory Invoice Header
PK,FK2 Invoice Number
Quantity Quantity PK Invoice Number PK Line Number

Invoice Date FK1 Item Number


FK1 Customer Number Quantity
Facts Unit Price

Invoice Invoice
Measures Quantity Quantity

Invoice
Revenue

Invoice Invoice
Unit Price Unit Price

© 2007 – Z Y Solutions Corporation

Designing your Data Warehouse

• What information do I want to derive


– Design to business needs, NOT just a replicate of the data
source
– Consider performance, de-normalize where appropriate

© 2007 – Z Y Solutions Corporation


Data Warehouse Schema Format

• Third Normal Form (3NF)


– Classical RDBMS
modeling technique
– Minimizes data
redundancy
through
normalization

© 2007 – Z Y Solutions Corporation

Data Warehouse Schema Format

• Star Schema Customer


Item

– Consists of a PK Customer Number


PK Item Number

single Fact table Customer Name


Customer Address
Item Description

– Compound
primary key Sales

PK,FK1 Customer Number


– One segment for PK,FK2
PK,FK3
Item Number
Territory Number
each Dimension PK,FK4 Warehouse Number

Quantity
Unit Price

Territory Warehouse

PK Territory Number PK Warehouse Number

© 2007 – Z Y Solutions Corporation

Data Warehouse Schema Format

• Snowflake Schema
– Variation on the
star schema
– Very large
dimension tables
are normalized
into multiple
tables

© 2007 – Z Y Solutions Corporation


Extract, Transform and Load (ETL)

• Extract
– Data extraction and staging
• Minimize impact on production data sources
• Transform
– Convert to format required by data warehouse
• Cleanse data to ensure accuracy
• Validate primary keys against defined owner
• Convert to different numbering schema
• Load
– Load data to data warehouse
• Follow guidelines as outlined by data warehouse

© 2007 – Z Y Solutions Corporation

ETL Tools

• Vendor tools
– Informatica
– DataMirror
– SAS ETL Solutions
• Microsoft
– SQL Server 2000
• Data Transformation Services (DTS)
– SQL Server 2005
• SQL Server Integration Services (SSIS)

© 2007 – Z Y Solutions Corporation

BI Tool vs. Query Tool

What’s the Difference?


Why Should I Care?

© 2007 – Z Y Solutions Corporation


B.I. Tool vs. Query Tool

• Knowledge of Database Structure Not Required


– Knowledge of Table Names & Relationships Not Needed
• Ability to Predefine Formulas
– Define once for everyone
• Multi-Pass SQL
– Produces Results Not Otherwise Possible
– Improve performance for complex queries

© 2007 – Z Y Solutions Corporation

B.I. Tool vs. Query Tool

• Single Version of the Truth


– Empower end users to create their own reports without
fear of varying results
• Control who sees what
– Row level security
– Object level security (including columns)
– Write one report to facilitate multiple user groups
• Manage and monitor usage

© 2007 – Z Y Solutions Corporation

Deployment Objectives

• Ease of administration
– One metadata layer definition shared by all 5 styles of BI
– One administrative interface for all 5 styles of BI
• Ease of deployment
– Zero foot print web interface
• Ease of use
– User friendly / intuitive interface. Web, drag and drop,
context sensitive menus, user definable help text
– Business terms familiar to the organization deploying
solution

© 2007 – Z Y Solutions Corporation


Deployment Objectives

• Reduce I.T. involvement


– Empower end-users to maintain existing and create new
reports
– Empower end users to schedule execution of reports and
select delivery method
• Web
• E-mail
• External Device
• Give end-users immediate access to data
• Allow end-users to mine for new data

© 2007 – Z Y Solutions Corporation

Full Feature BI Tool Example

MicroStrategy

© 2007 – Z Y Solutions Corporation

5 Styles of B.I. within 1 Unified Architecture

Web Browsers Enterprise Microsoft Office


Portals

Unified Web Interface


Alerts & Proactive
& Dashboards

Notification
Scorecards

Advanced
Reporting

Analysis
OLAP

Integrated Backplane
Services Oriented Architecture

Operational Operational Data Data SAP


Database Database Warehouse Mart BW
(ERP) (CRM, RFID)

Operational Databases Decision Support Databases

© 2007 – Z Y Solutions Corporation


Reusable Metadata

Other BI
Range of Metadata MicroStrategy Technologies

REPORT DESIGN Report-


Layout Specific
Format
Calculations Design
REPORT COMPONENTS
Parameterization Report-
Templates Specific
Filters
Autostyles
Design
BUSINESS ABSTRACTION
Metrics Reusable
Hierarchies
Custom Groupings Components
Transformations

DATA ABSTRACTION
Attributes Reusable
Facts Components
Tables
Aliases

© 2007 – Z Y Solutions Corporation

Development Efficiency

© 2007 – Z Y Solutions Corporation

MicroStrategy Architecture

Email, Wireless, Browsers


Portal
Print & Fax

MicroStrategy Office
MicroStrategy
Web

MicroStrategy Desktop MicroStrategy Desktop


MicroStrategy Architect
MicroStrategy Administrator

Web Services
MicroStrategy Intelligence Server
HTTP MicroStrategy OLAP Services
MicroStrategy Report Services
TCP/IP MicroStrategy Narrowcast Server

ODBC
SMTP/SMS
/Portal Metadata Warehouse

© 2007 – Z Y Solutions Corporation


MicroStrategy Architecture

Web Browser
Printer Web Browser Web Browser
Fax

HTTP HTTP
HTTP

Mobile
SMTP/
E-Mail
SMS/
Portal Web
Services
SMTP/
SMS/
MicroStrategy Desktop Portal Web
MicroStrategy Architect Services
MicroStrategy Administrator
Web
Services

Web
TCP/IP
TCP/IP Server
MicroStrategy Desktop MicroStrategy Intelligence Server
MicroStrategy OLAP Services
MicroStrategy Report Services
ODBC
MicroStrategy Narrowcast Server
e
ta
us
are Da
ho

ODBC
W

ta

MicroStrategy
da
eta

Intelligence
M

Server

MicroStrategy Architecture

© 2007 – Z Y Solutions Corporation

Interactive Demonstration

© 2007 – Z Y Solutions Corporation

Questions & Answers

Tom Peters
tpeters@zysolutions.com
(847) 487-2300, x204

© 2007 – Z Y Solutions Corporation

Das könnte Ihnen auch gefallen