Introduction to
Data Warehousing
From DBMS to Decision Support
• DBMSs widely used to maintain transactional
data
• Attempts to use of these data for analysis,
exploration, identification of trends etc. has led
to Decision Support Systems.
• Rapid Growth since mid 70’s
• DBMSs vendors have answered this trend by
adding new features to existing products
• Rarely enough
DBs for Decision Support
• Trend towards Data Warehousing
• Data Warehousing – consolidation of data
from several databases which are in turn
maintained by individual business units
along with historical and summary
information
Characteristics of TPSs
Characteristic OLTP
Typical operation Update
Level of analytical requirements Low
Screens Unchanging
Amount of data per transaction Small
Data level Detailed
Age of data Current
Orientation Records
TPS
TPS vs
vs Decision
Decision Support
Support
OLTP Complex Analysis
Information to support Historical information
day-to-day service to analyze
Data stored at transaction Data needs to be integrated
level
Database design: Normalized Database design:
Denormalized, star schema
MIS and Decision Support
Ad hoc access
Production
platforms
Operational reports Decision makers
• MIS systems provided business data
• Reports were developed on request
• Reports provided little analysis capability
• no personal ad hoc access to data
Analyzing Data from
Operational Systems
• Data structures are complex
• Systems are designed for high performance and
ERP
throughput
• Data is not meaningfully represented
• Data is dispersed
• TPS systems unsuitable for intensive queries
Production
platforms
Operational reports
Data Extract Processing
Operational systems Extracts Decision makers
• End user computing offloaded from
the operational environment
• User’s own data
Management Issues
Operational systems Extracts Decision makers
Extract explosion
• Duplicated effort
• Multiple technologies
• Obsolete reports
• No metadata
Data Quality Issues
• No common time basis
• Different calculation algorithms
• Different levels of extraction
• Different levels of granularity
• Different data field names
• Different data field meanings
• Missing information
• No data correction rules
• No drill-down capability
From Extract to Warehouse DSS
Internal and Data warehouse Decision makers
external systems
• Controlled
• Reliable
• Quality information
• Single source of
data
Data Warehousing Architecture
External Data Sources
Visualisation
Extract Clean
Metadata
Transform Load respository Serves
OLAP
Refresh
Operational Databases Data Warehouse
Data Mining
Business Motivators
• Provide superior services and
products
• Know the business
• New products
• Invest in customers
• Retain customers
• Invest in technology
• Reinvent to face new challenges
Centralised data warehouse
Corporate
Analyst
data-
warehouse
Mainframe Server Analyst
Corporate
Financial
Marketing
Analyst
Manufacturing
Distribution
Federated data warehouse
Financial Analyst
Corporate Marketing Analyst
data
Mainframe warehouse Manufacturing Analyst
Distribution Analyst
Tiered data warehouse
Mainframe
Tier 3 (detailed data) Corporate data warehouse
Tier 2 (summarized data) Local data mart
Tier 1 (highly summarized data) Workstation Analyst
Data Warehouses Vs Data
Marts
Data
Data Mart
Warehouse
Property Data Warehouse Data Mart
Scope Enterprise Department
Subjects Multiple Single-subject
Data Source Many Few
Size (typical) 100 GB to > 1 TB < 100 GB
Implementation time Months to years Months
End-user Access Tools
• High performance is achieved by pre-planning the
requirements for joins, summations, and periodic
reports by end-users.
• There are five main groups of access tools:
– Data reporting and query tools
– Application development tools
– Executive information system (EIS) tools
– Online analytical processing (OLAP) tools
– Data mining tools
Data Usage - $1000 questions
Verification Discovery
What is the average sale for What is the best predictor
in-store and catalog of sales?
customers?
What is the average high What are the best
school GPA of students who predictors of college
graduate from college graduation?
compared to those who do
not?
Need to complement RDBMS technology with a flexible,
multidimensional view of data
The Functionality of OLAP
• Rotate and drill down
• Create and examine calculated data
• Determine comparative or relative
differences.
• Perform exception and trend analysis.
• Perform advanced analytical functions
The star structure
Product Region
Model Nation
Type Facts District
Color Product Dealer
Region
Time
Channel
Revenue
Channel Expenses Time
Units Week
Year
Multidimensional Database
Customer Store Model Store
Time Time
SALES FINANCE
Product
The data is found at the intersection of
dimensions.
Data Mining
Data mining functions
• Associations
– 85 percent of customers who buy a certain brand of wine also buy a
certain type of pasta
• Sequential patterns
– 32 percent of female customers who order a red jacket within six months
buy a gray skirt
• Classifying
– Frequent customers are those with incomes about $50,000 and having
two or more children
• Clustering
– Market segmentation
• Predicting
– predict the revenue value of a new customer based on that personal
demographic variables
