Lecture 1

Mr. Suleiman M Yussuf

Before the 1970s Computers were thought to

provide only computational power
During the 70s, people started expecting
computers to also generate and maintain
processes that support the decision-making
capability of human beings
The primary reason for this change was the
information overload problem
Lots of data was starting to get generated on
a daily basis: Media, Internet, Enterprise
transactions etc.
Also, the development of the relational data
model facilitated database storage and retrieval.
2

High-level Enterprise employees needed to sift

through voluminous data in order to extract
useful information for making decisions
E.g., a 1-year record of sales for a particular
product, a frequency distribution of customers
for the past 6 months etc.
Not all the managerial needs could be satisfied
through traditional systems
Decision Support Systems (DSS) were developed
that specially prepared, separated, and staged
the data that was specifically needed for decision
support
Easy access to the needed data
Improves system response time
Enhances data integrity and security.
3

DSS Output example: Comparative sales and

Projected Revenue figures
Drawback: Managers couldnt operate DSSs
autonomously
Executive Information Systems (EIS): More
simplified systems in which the manager
instantly knew what was happening
Emphasis is on graphical displays and easyto-use user interfaces
Financials, production history, current
application status, plans, external events
(competitor information, emails etc.)
Being referred to as Business Intelligence.
4

Pros:
Ease of use for upper-level executives
Provides timely and efficient delivery
Information can be better understood
Cons:
System may become slow and large
High implementation costs
Limited functionality
Less reliability and security
Initial EISs didnt have analytical capability of
DSSs: both are used in conjunction
EISs are used to find problems, while DSSs
study them and offer alternative solutions.
5

Gate Assignment Display System (GADS)

Developed by Texas Instruments in 1987
for United Airlines.
Significantly reduced travel delays by aiding
the management of ground operations at
various airports
Gate Assignment is complex: Security, Gate
Capacity, some airplanes will fit at only some
gates etc.
Knowledge-based system (technique of AI) that
stores past data in order to reason from it and
generate decision-support information.
6

Clinical decision support systems for medical

diagnosis (Google for more details)
MYCIN: Diagnoses of bacterial diseases
CADUCEUS: Diagnoses 1000 diseases.
Iliad: Uses Bayesian reasoning to diagnose 1500
diseases
Lifecom: Tracks, processes and automatically
presents all relevant clinical considerations to the
physician, nurse practitioner, physician's assistant,
nurse, or medical assistant at the exact moment that
the knowledge can do the most good
RODIA: Relative Optical Density Image Analysis used
in medical imaging, medical diagnostics, orthopedic
and other medical disciplines.

Lack of a strong Database component for most

EISs and DSSs
Online Transaction Processing (OLTP): the
technology that facilitates and manages
transaction-oriented applications, e.g., ATMs
Database or Business transactions
Initial OLTP Databases: The stored
organizational information was directed to
maintaining current (i.e., online) information
about individual transactions and customers
Managerial information requires past as well as
future information
Companies developed their own Databases, but
suffered from a lack of technique and resources.
8

Typically used to capture new data or update

existing data, specifically in high throughput,
insert/update-intensive systems
Order-entry, airline reservation, ATM etc
Characteristics:
Transactions that involve small amounts of
data
Indexed and fast access to data
Many users and Frequent queries
OLTP systems support transactions that span
a network and may include more than one
company
Transaction Servers of Microsoft and IBM
Database query optimization techniques.

Pros:
Simplicity: Reduced paper trails
Efficiency: Faster, more accurate forecasts
for revenues and expenses
Cons:
Security: Online transaction systems are
generally more susceptible to direct attack
and abuse than their offline counterparts
Reliability: Operations can be severely
impacted if the database is unavailable due
to data corruption, systems failure, or
network availability issues
Cost: Offline maintenance is difficult

10

Developing and maintaining databases was

becoming a problem for each individual
enterprise
The Databases developed concentrated on
storing only online information
Only analysis of the current situation was
available to the managers
A wider management scope called for the
storage of past and future information
The introduction of new software development
methods/applications called for shorter
decision support life cycles (as compared to
EISs).
11

1990s: Organizations started developing data

warehouses to serve the decision support needs
Different from traditional systems
Use of special-purpose software that facilitates
the extraction, cleaning, and loading of data
(multi-dimensional databases, variety of
server softwares etc.)
Easy storage of past, present and future data
Enhanced data access tools facilitate the
autonomous access, analysis and display of
decision-support information (e.g., without
using SQL)
Time-based decision support (support when
required).
12

Initial Databases were designed specifically for

operational (transactional) use based on OLTP
In mid 1980s Database developers realized that
Storing historical data was important for the
managers
Complicated (analysis-based) queries could
hang up large transactional databases, thus
slowing the response and decision-making
times
Online Analytical Processing (OLAP) databases
were developed specifically for analysis
OLAP is at the heart of data warehouses.
13

An approach to quickly answer multidimensional analytical queries

Multidimensional Databases uses a variation of
the relational model that exploit
multidimensional structures to organize data and
express the relationships between data
First standard API: OLE DB for OLAP
Introduced the MDX query language
Second API: XML for Analysis (XMLA)
Business reporting for sales, marketing,
management reporting , business process
management (BPM), budgeting and forecasting,
financial reporting.

14