Data Mining

Lecture 2
Business Intelligence
Data Mining
Big Data
UCD Quinn School of Business

Scoil Gn Ui Chuinn UCD
Figure 2.1 Levels of managerial decision taking

Slide 2
BUSINESS INTELLIGENCE
Business intelligence
knowledge about customers,

competitors, suppliers, your own
internal operations, etc
Combined forms of information to
create real knowledge
Encompasses everything that
affects your business
Helps you make strategic business
decisions
Slide 3

McGraw-Hill
2007 The McGraw-Hill Companies, Inc. All rights reserved.
BUSINESS INTELLIGENCE
Slide 4

McGraw-Hill
2007 The McGraw-Hill Companies, Inc. All rights reserved.
Data Mining

What is Data Mining

Analyzing and summarizing
Analyze data from many different dimensions or
angles, categorize it, and summarize the
relationships identified.
Finding correlations or patterns

History of data mining

Data Mining saves Tesco 350 Million a

year
What scares me about this is that you know
more about my customers after three months
than I know after 30 years- Tescos then
Chairman Lord MacLaurin

The Urban Myth

Advertising on Facebook

Multidimensional Analysis
and Data Mining
Data mining the process of analyzing data to
extract information not offered by the raw data
alone
To perform data mining users need data-mining
tools
Data-mining tool uses a variety of techniques to find
patterns and relationships in large volumes of
information and infers rules that predict future behavior
and guide decision making

What Information do Businesses Need?

Need to convert accumulated Data into
useful Information
Timely, Accurate, and Consistent information and
analysis
Who are our best customers?
How can we retain them?
What are their preferences?
Who are our potential customers?
How can we develop and maintain those relationships?
What events lead to customer loss?
What are our problem areas?
Causes (Prevent future problems from occurring)
Solutions (Uncover previously unknown solutions)

Slide 12
Current State of E-Commerce

The Need For Business Intelligence & Data Mining
Investments in network infrastructure, hardware, and
operational software applications have provided a flood of
customer data.
Transaction-generated, automated data on
Customers
Partners
Services
Suppliers
Performance History

Inventory Levels
Products
Slide 13
Sources of Customer Data

Drowning in Customer Data from
Web Site
Registration
Name
Gender
Address
E-Mail Address
Customers
Daytime Phone
Evening Phone
Employer
Job Title
Income Level
Hobbies
Memberships

Interests
Purchases
Slide 14

Point of Purchase
Name
Gender
Address
E-Mail Address
Customers
Daytime Phone
Evening Phone
Employer
Job Title
Income Level
Hobbies
Memberships

Interests
Purchases
Slide 15

Externally Purchased
Marketing Data
Name
Gender
Address
E-Mail Address
Customers
Daytime Phone
Evening Phone
Employer
Job Title
Income Level
Hobbies
Memberships

Interests
Purchases
Slide 16
Business Intelligence (BI)

Understanding The Past & Predicting The Future
A Value Solution
Answers To Critical Business Questions Are Buried In This Data
Need to convert into usable information

With
Speed
Data can be quickly gathered, organized, and formatted
And
Effectiveness
Information can then be delivered to the right people at the right time

Slide 17

Understanding The Past & Predicting The Future
A Value Solution
Critical Information Enhances Business Performance

Information
gathered through BI leads to a better understanding

of customer behavior, enhances operations, and optimizes
strategic decision-making
Customers (Contact Information, Interests, Buying Habits, etc.)
Contact Preferences (E-mail, Phone, Direct Mail)
Marketing (Current Plans, Necessary Adjustments, Future Plans)
Operational Issues (Staffing, Service Levels, Supplier Performance)


Data Collection
Transformation
Extracting customer data from various sources

Organizing data into an understandable and usable format
Application of Analytical & Data Mining Tools

Comprehensive explanation of historical data
What Happened?When?

Slide 19

Data Collection
Transformation
Extracting customer data from various sources

Organizing data into an understandable and usable format
Application of Analytical & Data Mining Tools

Comprehensive explanation of historical data
What Happened?When?
Predict future performance and trends based on this historical
data
Consumer Demand Forecasts, Staffing Needs, etc
Based on the data weve accumulated from similar customers,
is this set of potential customers a credit risk?
Intelligent Strategy Execution

Develop, implement, and follow a strategy based on your
predictions
Slide 20
Business Intelligence (Basic Process

Flow)
Demographic
Data
Psychographic
Data
Externally
Purchased
Collection
Sales
Data
Organize
Enterprise
Data
Load
Data Mining
Reporting
Testing
Slide 21
Business Intelligence (Comprehensive

Framework)
Demographic
Data
Sales
Data
Psychographic
Data
Extraction, Transformation, Load
Externally
Purchased
Enterprise
Data
Business-Critical Information Consolidated

Analysis Conducted
Decision Support Tools (Data Mining)
Reports Generated, Produced, and Delivered
Slide 22
Data Mining
In Summary, A Business Intelligence Tool
Discovers previously unknown patterns within vast amounts of data
Similar to querying a database (or databases)
Instead of gathering raw data, data mining retrieves information
Types of Data Mining
Hypothesis Testing
Undirected Data Mining
Directed Data Mining

Slide 23
Data Mining
Hypothesis Testing
Answers Questions
Example
Q: Does the age of an insurance agent matter when trying
to sign-up new customers to new policies?
A: The assumption might have been that older, more experienced
agents will have more success. But, the data may reveal that the age
difference between the agent and the policy holder is most important.
(Young agents have more success with young customers while older
agents have success with older customers.)
Slide 24
Data Mining
Letting algorithms find patterns in vast amounts of data
Types
Automatic Cluster Detection
Market Basket Analysis
Sequential Pattern Matching

Finds clumps of customers

who are similar because of
income level, buying habits,
geographic location, etc.
Slide 25
Data Mining
Types

Determines what types of

products (and services)
are purchased at the same
point in time
Slide 26
Data Mining
Types

Determines events that

occur in a particular
sequence in time
Slide 27
Data Mining
Directed Data Mining
Known as Predictive or Profiling Data Mining
Applying data from the past to a similar business situation in the future
Customer Churn Example
A customer that has left in the past is similar to one who will leave
in the future. So, gather data on lost customers in order to hopefully
decrease the likelihood that current customers will leave.
Marketing Example
Customers who responded to an advertisement or purchased a
product in the past are similar to those who will buy in the future.
Slide 28
Data Mining In Action

Source: Jupiter Research
Scoil
Gn Ui Chuinn UCD
Week 4/4 Slide 29
Following the Enron Collapse

Mined a database of Enrons
financial statistics looking for
signs that might have predicted
the companys collapse
In August 2001, Enron was

reporting record earnings
and its stock was trading at
between $34 and $47 per share.
However, its cash flow was
negative.

Scoil
Gn Ui Chuinn UCD
Week 4/4 Slide 30
StockDiagnostics.com
follows
Operational Cash
Flow Per Share
(OPS)
and thousands of
other data points
for every
publicly traded company
in the United States
Uses numerous
pieces of data from each
companys past
financial statements
to determine
the companys
current health and
future performance.

Scoil
Gn Ui Chuinn UCD
Week 4/4 Slide 31
VIP Customers
VIP CustomersIdentify customers who
are power buyers
The top 20% of your customers account
for 80% of total sales
Retailers must constantly communicate
with these customers to promote new
products
Inventory AdvantageRetailer orders
are placed 6 months prior to a shopping
season. They first contact VIP customers
to buy new products (at a discount) then
adjust inventory levels based on their
response

Slide 32
BIG Data
Problems: too much
data
Need to establish
relationships and
patterns
Big data is the
solution not the
problem

Example: Carr: The Big Switch

Sunday, April 10, 20

16
Carr: A Spiders Web

Who is 4417749?

Search Logs
In July 2007, AOL released the keywords
entered into its search engine by 657.000
subscribers
To protect subscribers privacy, AOL had
anonymised the data, removing
identities

Who is 4417749?
On August 9, 2007 Themla Arnold, a 62
year old widow from Georgia, woke up to
find her picture on the national edition of
The Times

Through the search we

make, the sites we visit, we
disclose details about our
jobs, hobbies, families,
politics, health.
Our secrets, fantasies,
obsessions & even our
crimes
Our sense of anonymity is
largely an illusion

Who is 4417749?
But the detailed records
of searches conducted by
Ms. Arnold and 657,000
other Americans, copies
of which continue to
circulate online,
underscore how much
people unintentionally
reveal about themselves
when they use search
engines and how risky
it can be for companies
like AOL, Google and
Yahoo to compile such
data.
On the Internet, nobody knows you're a

dog

In reality, not only is it known that you are

a dog, but what breed you are, your age,
where you live, and the kind of treat your
prefer

As computer scientists continue to refine

data-mining algorithms, they are
uncovering new ways to predict how
people will react when they are
presented with information
They are learning not only how to
identify us but also how to manipulate us
The Internet puts

enormous power into
the hand of individuals,
but even greater power
into the hands of
governments, and
other institutions
whose business it is to
control individuals


Data Mining

Uploaded by

Copyright:

Available Formats

Data Mining

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Data Mining

Uploaded by

Copyright:

Available Formats

Lecture 2

UCD Quinn School of Business

Figure 2.1 Levels of managerial decision taking

knowledge about customers,

Scoil Gn Ui Chuinn UCD

2007 The McGraw-Hill Companies, Inc. All rights reserved.

UCD Quinn School of Business

Scoil Gn Ui Chuinn UCD

2007 The McGraw-Hill Companies, Inc. All rights reserved.

UCD Quinn School of Business

What is Data Mining

UCD Quinn School of Business

History of data mining

UCD Quinn School of Business

Data Mining saves Tesco 350 Million a

UCD Quinn School of Business

The Urban Myth

UCD Quinn School of Business

UCD Quinn School of Business

UCD Quinn School of Business

What Information do Businesses Need?

UCD Quinn School of Business

Current State of E-Commerce

UCD Quinn School of Business

Sources of Customer Data

UCD Quinn School of Business

Sources of Customer Data

Drowning in Customer Data from

UCD Quinn School of Business

Sources of Customer Data

UCD Quinn School of Business

Business Intelligence (BI)

Answers To Critical Business Questions Are Buried In This Data

Need to convert into usable information

Data can be quickly gathered, organized, and formatted

UCD Quinn School of Business

Business Intelligence (BI)

Critical Information Enhances Business Performance

gathered through BI leads to a better understanding

Customers (Contact Information, Interests, Buying Habits, etc.)

Contact Preferences (E-mail, Phone, Direct Mail)

Marketing (Current Plans, Necessary Adjustments, Future Plans)

Operational Issues (Staffing, Service Levels, Supplier Performance)

Business Intelligence (BI)

Extracting customer data from various sources

Application of Analytical & Data Mining Tools

UCD Quinn School of Business

Business Intelligence (BI)

Extracting customer data from various sources

Application of Analytical & Data Mining Tools

Intelligent Strategy Execution

Business Intelligence (Basic Process

Business Intelligence (Comprehensive

Extraction, Transformation, Load

Business-Critical Information Consolidated

UCD Quinn School of Business

Scoil Gn Ui Chuinn UCD

Discovers previously unknown patterns within vast amounts of data

Similar to querying a database (or databases)