Download ch18

Business Intelligence
Transparencies
1
Objectives
 What business intelligence (BI) represents.
 The technologies associated with business intelligence
including: data warehousing, online analytical
processing (OLAP), and data mining.
 The main concepts associated with a data warehouse.
 The relationship between online transaction
processing (OLTP) systems and a data warehouse.
 The main concepts associated with a data mart.
©Pearson Education 2009
2
Objectives
 Designing a database for decision-support using a
technique called dimensionality modeling.
 The important concepts associated with online
analytical processing (OLAP) systems.
 The main categories of OLAP tools.
 The main concepts associated with data mining.
 How a business intelligence (BI) tool such as
Microsoft Analytical Services provides decisionsupport.
©Pearson Education 2009
3
Business intelligence
 The processes for collecting and analyzing data, the
technologies used in these processes, and the
information obtained from these processes with the
purpose of facilitating corporate decision–making.
 The main technologies associated with business
intelligence includes:
 data warehouse,
 online analytical processing (OLAP),
 data mining.
©Pearson Education 2009
4
Data warehouse
 A database system that is designed to support
decision-making by presenting an integrated view
of corporate data that is copied from disparate data
sources.
 Data held in a data warehouse is described as
being subject-oriented, integrated, time-variant,
and non-volatile (Inmon, 1993).
 The main source of data for the data warehouse are
online transaction processing (OLTP) systems.
©Pearson Education 2009
5
Comparison of OLTP with data
warehousing
©Pearson Education 2009
6
Typical architecture of a data
warehouse
©Pearson Education 2009
7
Data mart
 A subset of a data warehouse, which supports the
decision-making requirements of a particular
department or business area.
©Pearson Education 2009
8
Designing databases for
decision-support
 Decision-support databases can be designed using
traditional database design or specialist techniques
such as dimensionality modeling.
 Dimensionality modeling aims to build a data model
(called dimensional model) that has a consistent and
intuitive structure to facilitate efficient multidimensional analysis of data.
©Pearson Education 2009
9
Dimensionality modeling
 Creates a dimensional model (DM) called a star
schema that has a fact table containing factual data in
the center, surrounded by smaller dimension tables
containing denormalized reference data.
 As the bulk of data is represented as facts, the fact
tables can be extremely large relative to the dimension
tables.
 Dimension tables contain descriptive textual
information and are used as the constraints (search
conditions) in queries on the fact data.
©Pearson Education 2009
10
Star schema for StayHome DVD
sales
©Pearson Education 2009
11
Online analytical processing (OLAP)
 Stores large volumes of multi-dimensional data
that is aggregated (summarized) to various
levels of detail to support advanced analysis of
this data.
 Multi-dimensional data can be characterized
through many different views. For example DVD
sales can be viewed by product, customer, and/or
sales channel.
©Pearson Education 2009
12
Examples of OLAP applications
©Pearson Education 2009
13
Online analytical processing (OLAP)
 OLAP tools are categorized according to the
architecture of the underlying database (providing
the data for the purposes of online analytical
processing).
 There are three main categories of OLAP tools:
 Multi-dimensional OLAP (MOLAP or MD-OLAP);
 Relational OLAP (ROLAP);
 Hybrid OLAP (HOLAP).
©Pearson Education 2009
14
Typical architecture for multidimensional OLAP (MOLAP)
©Pearson Education 2009
15
Typical architecture for relational
OLAP (ROLAP)
©Pearson Education 2009
16
Typical architecture for hybrid OLAP
(HOLAP)
©Pearson Education 2009
17
Cube Browser of Microsoft SQL
analytical services
©Pearson Education 2009
18
Data mining
 The process of extracting valid, previously
unknown, comprehensible, and actionable
knowledge from large databases and using it to
provide decision-support.
©Pearson Education 2009
19
Examples of data mining
applications
©Pearson Education 2009
20
Data mining tools
 Important features of data mining tools include;
 data preparation;
 selection of data mining operations (algorithms);
 product scalability and performance;
 facilities for understanding results.
©Pearson Education 2009
21
Data Mining Model Browser of
Microsoft SQL Analytical Services
©Pearson Education 2009
22
Dependency Network Browser of
Microsoft SQL Analytical Services
©Pearson Education 2009
23
Data warehousing and data mining
 Major challenge to exploit data mining is
identifying suitable data to mine.
 Data mining requires a single, separate, clean,
integrated, and self-consistent source of data.
 A data warehouse is well equipped for providing
data for mining.
©Pearson Education 2009
24