Data Warehousing

From Nagpur university syllabus 2nd

sem MBA IT

Ashish Vinchurney

For more detailed notes you can also refer my blog

http://rtmnupervasivecomp.blogspot.com/
Concept
A warehouse is a subject oriented, integrated, time
variant and non volatile collection of data in support of
management’s decision making process.
The main terms –
Subject Oriented – Information about particular subject
.
Integrated – Data gathered various sources into one
database.
Time Variant – Data identified with particular time
period.
Non Volatile – Data which is stable.
Concept …
Basically Data Warehouse is combining
data from multiple sources in one
comprehensive database.
Data warehouse includes queries, analysis
and reporting.
Importance Of Data Warehousing
Data Warehousing is needed for the
following purposes –

Building a comprehensive database

Making up to date database
Database analysis
View day to day operations
Strategic decision making
History ..
The Key developments in early years of
data warehousing are .
1960 – General Mills and Datrmouth
College in joint research project.
1970 – ACNielson and IRI provide
dimensional data marts for retail sales.
1983 – Teradata introduces a database
management sys specifically designed for
decision support .
History Contd…
1988 – Barry Devlin & Paul murphy published
the article An architecture for a business & info.
Sys . In IBM systems Journal where they
introduced the term “Business data warehouse”
1990 – Red Brick systems introduces Red Brick
Warehouse a database mang sys. for data
warehousing.
1991 – Prism solutions intorduced Prism
Warehouse manager s/w for developing a
warehouse.
History Contd…
1995 – The data warehousing institute
that promotes data warehousing is
founded
1996 – Ralph kimball publishes the book
The Data Warehouse Toolkit
1997 – Oracle 8 with support for star
queries is released.
Building a Warehouse
Storage methods
Architecture
Developing Strategy
Maintenance issues
Storage Methods
Data Warehouse have various schemas
designed for data analysis. The data is
usually multidimensional data with different
attributes.

The multidimensional data is stored in tables

called fact tables.

To minimize the storage requirements the

dimension attributes are used which act as
foreign keys into other tables called
dimension tables.
The complex schema formed as a result of
the use of dimension attributes for
referencing the dimension tables is called
as star schema.

Warehouse may have many levels of

dimensions . If multiple levels are there in
the warehouse then it is called snowflake
schema.
Architecture
Architecture is overall conceptualization
how data warehouse is to be build.
One possible conceptualization having
interconnected layers is as given below.
1) Operational layer
2) Data access layer
3) metadata layer
4) Informational access layer
Architecture Explained…
• Operational database layer – The source
data for warehouse , like an ERP system .
• Data access layer – Interface between
operational and informational access layer
containing tools to extract, transform, load
data.
• Metadata layer – The data directory means
the data about the data.
• Informational access layer – The data
accessed for reporting and analyzing and
tools for reporting and analyzing data. Like
OLAP tools.
 The data Warehouse Architecture

Data source 1
Data
loaders
Data source 2

Data source n DBMS

Query and
Data Warehouse analysis tools
Strategy / Issues involved in
development

When and how to gather data

o Source driven
o Destination driven
What schema to use
Data transformation
Propagating updates
Maintenance issues
For maintenance of the data warehouse these steps
should be followed .
 Train the users one step at a time
 Look closely at the data extracting, cleaning, and
loading tools
 Implement a user accessible automated directory to
information stored in the warehouse
 Determine a plan to test the integrity of the data in the
warehouse
Application In Strategic Decision
making
The data warehouse finds its use in
various places where the prediction is
needed for making decisions .
Banking & finance – To see the credit
worthiness of any customer .
Defense – Used against terrorist
intelligence.
Consumer Goods –

 Which are our lowest/highest margin

customer
 Who are my customers and what products
are they buying
 Which customers are most likely to go
to the competition
 Analysis of market
Web Data Analysis
Concept :-
 Web data which is available on net is very eccentric. An
expert user in data analysis is able to get most of it. But the
naive user may be lost.
 Web data analysis helps the companies to improve their
web sites by modifying the design which gives most of the
information.
 Even a little change in the content on the web site can make
a big difference.
 Because the web sites change so much the value of data
looses.
 Web data analysis can be profitable but it has to be
used intelligently.
 Web data does not give you much information about
the web site user.
 The companies have to know who are the real users
acknowledge the data problems and how can they be
benefitted by the data which is relevant.
For more detailed notes you can also refer my blog

http://rtmnupervasivecomp.blogspot.com/