Transcript Data

MP3 / MD740
Strategy & Information Systems
Oct. 13, 2004
Databases & the Data Asset, Types of
Information Systems, Artificial Intelligence
Topics Covered
• Data & Information
• Data Warehousing
• Leveraging Data
– Harrah’s
– Business Intelligence
• Types of Information Systems
– TPS, MIS, DSS
• Artificial Intelligence
– Expert Systems, Neural Networks, Genetic
Algorithms
Data, Information, & Knowledge
• Data - raw facts, figures, and details.
• Information - organized, meaningful, and
useful interpretation of data. Has a context,
answers a question.
• Knowledge - an awareness and
understanding of a set of information and
how that information can be put to best use.
• Many firms are data rich and info poor:
victims of an old or poorly planned
architecture
Examples of Data, Information, &
Knowledge
Data: raw, no context
900,000
1,200,000
1,150,000
1,100,000
Information: meaningful, has context
Quarter 1 Quarter 2
Post
900,000 1,150,000
Kellogg's 1,200,000 1,100,000
Knowledge: information above & other information
creates an awareness of impact
Post lowered its prices after the first quarter.
Price change has caused Post sales to rise at the expense
of Kellogg’s
Warehouses & Marts
• Data Warehouse
– a database designed to support decision-making in
an organization. It is structured for fast online
queries and exploration. Data warehouses may
aggregate enormous amounts of data from many
different operational systems.
• Data Mart
– a database focused on addressing the concerns of
a specific problem or business unit (e.g. Marketing,
Engineering). Size doesn’t define data marts, but
they tend to be smaller than data warehouses.
Data Warehouses & Data Marts
3rd party data
Data Mart
(Marketing)
TPS
& other
operational
systems
Data
Warehouse
= operational clients
= query, OLAP, mining, etc.
Data Mart
(Engineering)
Differing System Demands
Operational Systems
network traffic
& processor
demands
time
Managerial Systems
network traffic
& processor
demands
time
“Let the neighbors lure tourists with knights
on horseback, fiery volcanoes, pirate ships,
and mini-Manhattans. We’ll just keep
refining what we’re already pretty good at:
drilling into our data and making sure our
regular customers are more than satisfied.”
- Gary Loveman, CEO, Harrah’s
Types/Classifications of
Information Systems
Transaction Processing Systems
(TPS)
• A shared IS that uses a combination of IT and
manual procedures to process data and
information and to manage transactions.
• Examples: Cash Registers (POS), ATM
• Characteristics:
– transactions are similar & repeatable
– support multiple users in routine, everyday
transactions (usually tactical systems)
– data capture with possible report generation
– accuracy is critical, TPS “feed” other IS
Reporting Systems - MIS
• Sometimes called Management Reporting
Systems or Management Information
Systems
• Characteristics
– use data captured and stored from TPS
– reports consolidated information rather than
details of transactions
– supports reoccurring decisions
– provides reports in pre-specified formats (on
screen, printed, or data)
Decision Support Systems (DSS)
• Allow users interrogate computers on an ad hoc
basis, analyze information, and predict the
impact of decisions before they are made. [key:
unstructured, user-led exploration]
• Characteristics
– Assists in ad-hoc decision making
– Used when requirements, processes, or procedures
are unstructured & aren’t known in advance
– Provides info needed to define & solve a problem
– Provides information in format determined at time of
need
Management Levels & IS
Strategic
Planning
Management
Control
Operational
Control
DSS
MIS
TPS
Expert Systems (ES)
• An artificial intelligence system that uses
captured human expertise to evaluate and
solve problems
• Characteristics:
– diagnosis, configuration, and/or recommend a
course of action
– problems are structured and repeatable
– application scope is limited to a particular
problem area (domain)
Other Types of Artificial
Intelligence (AI)
• Neural Networks
– hunt for patterns in historical data
– build their own expertise based on prior history
– require clean data & consistency between
performance history and future events
• Genetic Algorithms
– search for optimal solutions based on natural
selection: (1) propose solution (2) evaluate
results against earlier solution (3) mutate &
return to step 1
Keane’s Space Truss Design