Transcript Chapter 9

Chapter 9
Business Intelligence and
Knowledge Management
Agenda
•
•
•
•
•
•
Business Intelligence System
Reporting system
Data Warehouse
Data Mart
Knowledge Management Systems
Discussion and Case Study
Business Intelligence System
• Need
– Inexpensive storage
– Drowning in data (terabyte - 12, petabyte - 15,
exabyte - 18)
– Starving for useful information
• Purpose
– Provide the right information, to the right user, at the
right time for actions
• Business intelligence tool
– Searching business data for finding patterns
– Types: reporting tool and data-mining tool
Reporting Tool
• Programs
– Read data from sources
– Sort and group data
– Calculate simple totals and averages
– Produce reports
– Deliver reports to the users
• For business assessment: a customer
canceling an important order
Data-mining Tool
• Programs
– Use sophisticated statistical techniques and
complex mathematics
– Search for patterns and relationships among
data
• For business prediction using probability
– Calculating the probability of a customer
defaulting on a loan
– Assessing new loan applications
Reporting System - I
• Purpose
– Create meaningful information from disparate data sources and
to deliver that information to the proper user on a timely basis
• Operation
–
–
–
–
Filtering data
Sorting data
Grouping data
Making simple calculations
• Component
– A database of reporting metadata with description of reports,
users, groups, roles, events, and other entities in the reporting
activity
Reporting System - II
• Report type
–
–
–
–
Static
Dynamic
Query
Online analytical process (dynamic grouping
structure)
• Report media
– Paper
– Voice
– Digital: screen, digital dashboard, Web service, email
alert
Reporting System - III
• Report mode
– Push: preset schedule
– Pull: user request
• Function
– Authoring: connecting to data sources, creating report
structure, and formatting report
– Management: who, what, when, by what mean, user
account, and user group
– Delivery: push or pull, method, time
• Example
– RFM analysis
– Online analytical processing (OLAP)
RFM Analysis
• Analyzing and ranking customers
according to their purchasing patterns
– How recently (R) a customer has ordered
– How frequently (F) a customer orders
– How much money (M) the customer spends
per order
RFM Score
• The program first sorts customer purchase records by
the date of their most recent (R) purchase
• The program then divides the customers into five groups
and gives customers in each group a score of 1 to 5.
– The top 20% of the customers having the most recent orders are
given an R score 1 (highest).
• The program then re-sorts the customers on the basis of
how frequently they order.
– The top 20% of the customers who order most frequently are
given a F score of 1 (highest).
• Finally the program sorts the customers again according
to the amount spent on their orders.
– The 20% who have ordered the most expensive items are given
an M score of 1 (highest).
OLAP
• Characteristics
– Provide the ability to sum, count, average, and other
simple arithmetic operations on groups of data
– Display the current state of the business
– The viewer can dynamically the report’s format
– Drill down (detail data)
• Component
– Measure: the data item of interest (total, average)
– Dimension: a characteristic of a measure (customer
type, sales region)
• OLAP server & OLAP database: store results
from operational databases
Role of OLAP Server and OLAP Database
Problems with Operational Data
• Problematic data (dirty data)
– Missing elements
– Inconsistent data
•
•
•
•
Nonintegrated data
Too fine or too coarse (clickstream data)
Wrong granularity (format)
Curse of dimensionality: the more attributes, the
easier to build a model to fit the sample data but
worthless as a predictor
Data Warehouse
• Programs read operational data and extract,
clean, and prepare data for business intelligence
processing
• Data-warehouse DBMS
– Extract and provide data to business intelligence tools
such as data-mining programs
– Internal data and purchased from outside sources
– Metadata: source, format, assumption, constraint, and
other facts about the data
Components of a Data Warehouse
Data Mart
• A data collection, smaller than the data
warehouse, to address a particular
component or functional area of the
business
• Expensive to create, staff, and operate
data warehouse and data mart
Data Mart Examples
Data Mining
• The application of statistical and
mathematic techniques to find patterns
and relationships among data for
classifying and predicting
• From artificial intelligence and machinelearning
• Type
– Unsupervised data mining
– Supervised data mining
Convergence Disciplines for Data Mining
Unsupervised Data Mining
• No model or hypothesis before running the analysis
• Apply the data-mining technique to the data and observe
the results
• Create hypotheses after the analysis to explain the
patterns found
• Cluster analysis
– Find groups of similar customers from customer order and
demographic data
• Decision Tree
– A hierarchical arrangement of criteria to predict a classification or
a value
– Loan-decision rules
Supervised Data Mining
• Develop a model prior to the analysis and apply
statistical techniques to data to estimate
parameters of the model
• Regression analysis
– Measure the impact of a set of variables on another
variable
• Neural network
– Predict values and make classifications such as
“good prospect” or “poor prospect” customers.
Market-Basket Analysis
• A data-mining technique for determining
sales patterns
• Show the products that customers tend to
buy together
• Support: the probability that two items will
be purchased together
• A standard CRM analysis
Knowledge Management (KM)
• The process of creating value from intellectual
capital and sharing that knowledge with
employees, managers, suppliers, customers,
and others
• Emphasis is on people, their knowledge, and
effective means for sharing that knowledge with
others
• Preserve organizational memory by capturing
and storing the lessons learned and best
practices of key employees
• Enable employees and others to leverage
organizational knowledge to work smarter
Benefits of KM
•
•
•
•
Free flow of ideas (innovation)
Storing lesson learned and best practice
Better customer service
Boosting profit by getting product to the
market faster
• Increasing employee retention
• Reducing cost by eliminating redundant
and unnecessary process
KM Content Management - I
• Track organizational documents, Web
pages, graphics, and related materials
• Concern with the creation, management,
and delivery of documents for a specific
KM purpose
KM Content Management -II
• Problems
– Complicated and huge
– Dependency relationship between documents
– Perishable document contents
– Multinational languages
• Delivering methods
– Pull using index and search engine
– Web browsers
Knowledge Sharing
• Portals, discussion groups, and email
– Idea publishing
– Bulletin board
– Frequent ask question
• Collaborations system
– Web broadcast
– Video conference
– Net meeting
• Expert system
– Decision tree with narrow domain and complex rules
– Expensive and difficult to create and maintain
Issues of Knowledge Sharing
• Problems
– Competition
– Shy
• Strategy
– Reward
– Incentive
Discussion
• Security (275a-b)
– State some methods for an organization to prevent the semantic
security problems.
• Problem Solving (283a-b)
– State two statistic usages and its associated risks in a business
decision making process.
• Ethics (289a-b)
– State some disadvantages of using decision tree as the
admission rules.
• Reflections (295a-b)
– Is it a common practice of lower management to manipulate the
data and generate the information to accommodate the needs of
upper management in the real business world? How do you
avoid this situation as the upper management?
Case Study
• Case 9-1 Laguna Tools (300-301) every
question
Points to Remember
•
•
•
•
•
•
Business Intelligence System
Reporting system
Data Warehouse
Data Mart
Knowledge Management Systems
Discussion and Case Study