DATA MINING

HAVE YOU NOTICED????????????

A call detail record is generated at the local phone company showing, among other
things, the time of your call, the number you dialed, and the long-distance company
to which you have been connected.
At the long-distance company, similar records are generated recording the duration
of your call and the exact routing it takes through the switching system.
This data will be combined with other records that store your billing plan, name, and
address in order to generate a bill.
At the catalog company, your call is logged again along with information about the
particular catalog from which you ordered and any special promotions you are
responding to.
When the customer service representative that answered your call asks for your
credit card number and expiration date, the information is immediately relayed to a
credit card verification system to approve the transaction; this too creates a record.
All too soon, the transaction reaches the bank that issued your credit card, where it
appears on your next monthly statement. When your order, with its item number, size,
and color, goes into the cataloger’s order entry system, it spawns still more records in
the billing system and the inventory control system. Within hours, your order is also
generating transaction records in a computer system
Phone companies have used call detail records to discover residential phone numbers
whose calling patterns resemble those of a business in order to market special
services to people operating businesses from their homes.
Catalog companies have used order histories to decide which customers should be
included in which future mailings
In fact, many companies gather hundreds of gigabytes or terabytes of data from and
about their customers
For learning to take place, data from many sources—billing records, scanner data,
registration forms, applications, call records, coupon redemptions, surveys—must first
be gathered together and organized in a consistent and useful way. This is called
data warehousing.
For the retail industry, a good example would be a retail data mart that incorporates
customer information from cash registers, mailing lists, websites, and feedback cards.
Similarly, another relevant example of the application is the healthcare sector that
uses it to access the patient’s reports, share important data with insurance providers,
predict outcomes, etc.
In healthcare, these central stores of data are used to record patient information
from different units of the medical unit. This would include patient personal
information, financial transactions with the hospital, and insurance data. All this is
consolidated in the data warehouse and connected through the database schema.
One of the most important aspects of the data warehouse is the capability to track
customer behavior over time.
Many of the patterns of interest for customer relationship management only become
apparent over time.
 Is usage trending up or down?
 How frequently does the customer return?
 Which channels does the customer prefer?
 Which promotions does the customer respond to?
The data warehouse provides the enterprise with a memory. But, memory is of little
use without intelligence. Intelligence allows us to comb through our memories, noticing
patterns, devising rules, coming up with new ideas, figuring out the right questions,
and making predictions about the future.
Who is likely to remain a loyal customer and who is likely to jump ship?
What products should be marketed to which prospects?
What determines whether a person will respond to a certain offer?
Which telemarketing script is best for this call?
Where should the next branch be located?
What is the next product or service this customer will want?
Answers to questions like these lie buried in corporate data.
DATA MINING
Data mining is the process of sorting through large data sets to identify patterns and
relationships that can help solve business problems through data analysis. Data mining
techniques and tools enable enterprises to predict future trends and make more-
informed business decisions.
Data mining is a step in the knowledge discovery in databases (KDD) process, a data
science methodology for gathering, processing and analyzing data.
“Data Mining is the exploration and analysis of large quantities of data in order to
discover meaningful patterns and rules. “
Data mining is largely concerned with building models.
A model is simply an algorithm or set of rules that connects a collection of inputs
(often in the form of fields in a corporate database) to a particular target or
outcome.
.Under the right circumstances, a model can result in insight by providing an
explanation of how outcomes of particular interest, such as placing an order or
failing to pay a bill, are related to and predicted by the available facts.
WHAT TASKS CAN BE PERFORMED WITH DATA
MINING?
■■ Classification
■■ Estimation
■■ Prediction
■■ Affinity grouping
■■ Clustering
■■ Description and profiling
CLASSIFICATION
Classification, one of the most common data mining tasks, seems to be a human
imperative. In order to understand and communicate about the world, we are
constantly classifying, categorizing, and grading.
Classification consists of examining the features of a newly presented object and
assigning it to one of a predefined set of classes.
■■ Classifying credit applicants as low, medium, or high risk
■■ Determining which phone numbers correspond to fax machines
■■ Spotting fraudulent insurance claims
■■ Assigning industry codes and job designations on the basis of free-text job
descriptions
ESTIMATION
Classification deals with discrete outcomes
Estimation deals with continuously valued outcomes. Given some input data, estimation
comes up with a value for some unknown continuous variable such as income, height,
or credit card balance.
■■ Estimating the number of children in a family
■■ Estimating a family’s total household income
■■ Estimating the lifetime value of a customer
Regression models and neural networks are well suited to estimation tasks
PREDICTION
Prediction is the same as classification or estimation, except that the records are
classified according to some predicted future behavior or estimated future value
The historical data is used to build a model that explains the current observed
behavior. When this model is applied to current inputs, the result is a prediction of
future behavior.
The primary reason for treating prediction as a separate task from classification and
estimation is that in predictive modeling there are additional issues regarding the
temporal relationship of the input variables or predictors to the target variable.
PREDICTION
Examples of prediction tasks addressed by the data mining techniques may be :
■■ Predicting the size of the balance that will be transferred if a credit card
prospect accepts a balance transfer offer
■■ Predicting which customers will leave within the next 6 months
■■ Predicting which telephone subscribers will order a value-added service such as
three-way calling or voice mail
The choice of technique depends on the nature of the input data, the type
of value to be predicted, and the importance attached to explicability of
the prediction.
AFFINITY GROUPING OR ASSOCIATION RULES
The task of affinity grouping is to determine which things go together. The
prototypical example is determining what things go together in a shopping cart at
the supermarket, the task at the heart of market basket analysis.

Retail chains can use affinity grouping to plan the arrangement of

items on store shelves or in a catalog so that items often purchased
together will be seen together.
CLUSTERING
Clustering is the task of segmenting a heterogeneous population into a number of
more homogeneous subgroups or clusters. What distinguishes clustering from
classification is that clustering does not rely on predefined classes. In classification,
each record is assigned a predefined class on the basis of a model developed
through training on preclassified examples.
PROFILING
Sometimes the purpose of data mining is simply to describe what is going on in a
complicated database in a way that increases our understanding of the people,
products, or processes that produced the data in the first place. A good enough
description of a behavior will often suggest an explanation for it as well
The promise of data mining is to find the interesting patterns lurking in all these
billions and trillions of bytes. Merely finding patterns is not enough. You must respond
to the patterns and act on them, ultimately turning data into information, information
into action, and action into value. This is the virtuous cycle of data mining in a nutshell.
THE VIRTUOUS CYCLE OF DATA MINING
Identifying the business problem.
Mining data to transform the data into actionable information.
Acting on the information.
Measuring the results.
IDENTIFY THE BUSINESS OPPORTUNITY
The virtuous cycle of data mining starts with identifying the right business
opportunities.
Unfortunately, there are too many good statisticians and competent analysts whose
work is essentially wasted because they are solving problems that don’t help the
business.
Good data miners want to avoid this situation
Many normal business processes are good candidates for data mining: Planning for a
new product introduction
■■ Planning direct marketing campaigns
■■ Understanding customer attrition/churn
■■ Evaluating results of a marketing test
Measurements of past efforts and ad hoc questions about the business also suggest
data mining opportunities:
■■ What types of customers responded to the last campaign?
■■ Where do the best customers live?
■■ Are long waits at automated tellers a cause of customers’ attrition?
■■ Do profitable customers use customer support?
DATA MINING

Data comes in many forms, in many formats, and from multiple systems, as shown in Figure 2.2.
Identifying the right data sources and bringing them together are critical success factors
Every data mining project has data issues:
inconsistent systems,
table keys that don’t match across databases,
records overwritten every few months, and so on.

“What can be done with available data?”

TAKING ACTION
Taking action is the purpose of the virtuous cycle of data mining. As already
mentioned, action can take many forms. Data mining makes business decisions more
informed. Over time, we expect that better-informed decisions lead to better results.
■■ Prioritizing customer service
■■ Adjusting inventory levels
The results of data mining need to feed into business processes that touch customers
and affect the customer relationship.
The importance of measuring results has already been highlighted. Despite its importance, it is
the stage in the virtuous cycle most likely to be overlooked. Even though the value of
measurement and continuous improvement is widely acknowledged, it is usually given less
attention than it deserves. How many business cases are implemented, with no one going back
to see how well reality matched the plans?
Individuals improve their own efforts by comparing and learning, by asking questions about
why plans match or do not match what really happened, by being willing to learn that earlier
assumptions were wrong. What works for individuals also works for organizations.
The time to start thinking about measurement is at the beginning when identifying the business
problem. How can results be measured?
A company that sends out coupons to encourage sales of their products will no doubt measure
the coupon redemption rate.
However, coupon-redeemers may have purchased the product anyway. Another appropriate
measure is increased sales in particular stores or regions, increases that can be tied to the
MEASURING ACTION
The importance of measuring results has already been highlighted. Despite its importance, it is
the stage in the virtuous cycle most likely to be overlooked. Even though the value of
measurement and continuous improvement is widely acknowledged, it is usually given less
attention than it deserves. How many business cases are implemented, with no one going back
to see how well reality matched the plans? Individuals improve their own efforts by comparing
and learning, by asking questions about why plans match or do not match what really
happened, by being willing to learn that earlier assumptions were wrong. What works for
individuals also works for organizations.
The time to start thinking about measurement is at the beginning when identifying the business
problem. How can results be measured?
A company that sends out coupons to encourage sales of their products will no doubt measure
the coupon redemption rate. However, coupon-redeemers may have purchased the product
anyway.
Another appropriate measure is increased sales in particular stores or regions, increases that
can be tied to the
DATA MINING METHODOLOGY
Data mining is all about creating models. Models take a set of inputs and produce an
output. The data used to create the model is called a model set. When models are
applied to new data, this is called the score set. The model set has three components,
which are discussed in more detail later in the chapter:
■■ The training set is used to build a set of models.
■■ The validation set1 is used to choose the best model of these.
■■ The test set is used to determine how the model performs on unseen data.
 Data mining techniques can be used to make three kinds of models for three kinds of
tasks: descriptive profiling, directed profiling, and prediction.

Input Model Output

Descriptive models describe what is in the data.
The output is one or more charts or numbers or graphics that explain what is going
on.
Hypothesis testing often produces descriptive models.
PROFILING
It needs not involve any sophisticated data analysis. Surveys, for instance, are one common
method of building customer profiles.
Surveys reveal what customers and prospects look like, or at least the way survey responders
answer questions.

Profiles are often based on demographic variables, such as geographic location, gender, and
age

Simple profiles are used to set insurance premiums.

A 17-year-old male pays more for car insurance than a 60-year-old female.
Similarly, the application form for a simple term life insurance policy asks about age, sex, and
smoking—and not much more.
PREDICTION

Profiling uses data from the past to describe what happened in the past.
Prediction goes one step further.
Prediction uses data from the past to predict what is likely to happen in the future.
This is a more powerful use of data.
Building a predictive model requires separation in time between the model inputs or
predictors and the model output, the thing to be predicted.
If this separation is not maintained, the model will not work
DIRECTED AND UNDIRECTED DATA MINING
The first three tasks, classification, estimation, and prediction are examples of
directed data mining. Affinity grouping and clustering are examples of undirected
data mining. Profiling may be either directed or undirected. In directed data mining
there is always a target variable—something to be classified, estimated, or
predicted. The process of building a classifier starts with a predefined set of classes
and examples of records that have already been correctly classified. Similarly, the
process of building an estimator starts with historical data where the values of the
target variable are already known. The modeling task is to find rules that explain the
known values of the target variable.
DIRECTED AND UNDIRECTED DATA MINING
In undirected data mining, there is no target variable.
The data mining task is to find overall patterns that are not tied to any one variable.
The most common form of undirected data mining is clustering, which finds groups of
similar records without any instructions about which variables should be considered as
most important.
Undirected data mining is descriptive by nature, so undirected data mining techniques
are often used for profiling, but directed techniques such as decision trees are also
very useful for building profiles.
In the machine learning literature, directed data mining is called supervised learning
and undirected data mining is called unsupervised learning.
DATA MINING METHODOLOGY
Translate the business problem into a data mining problem.
Select appropriate data.
Get to know the data.
Create a model set.
Fix problems with the data.
Transform data to bring information to the surface.
Build models.
Asses models.
Deploy models.
Assess results.
Begin again.
TRANSLATE THE BUSINESS PROBLEM INTO A DATA
MINING PROBLEM
General goals should be broken down into more specific ones to make it easier to
monitor progress in achieving them.
Gaining insight into customer behaviour might turn into concrete goals:
■■ Identify customers who are unlikely to renew their subscriptions.
STEP TWO: SELECT APPROPRIATE DATA

What Is Available?
The first place to look for data is in the corporate data warehouse. Data in the warehouse has
already been cleaned and verified and brought together from multiple sources. A single data
model hopefully ensures that similarly named fields have the same meaning and compatible
data types throughout the database
Unfortunately, there is no simple answer to this question. The answer depends on the particular
algorithms employed, the complexity of the data, and the relative frequency of possible
outcomes. Statisticians have spent years developing tests for determining the smallest model
set that can be used to produce a model.
Machine learning researchers have spent much time and energy devising ways to let parts of
the training set be reused for validation and test.
Data mining is most useful when the sheer volume of data obscures patterns that might be
detectable in smaller databases
How Much History Is Required
how far in the past should the data come from?
This is another simple question without a simple answer. The first thing to consider is seasonality
data from too far in the past may not be useful for mining because of changing market
conditions. This is especially true when some external event such as a change in the regulatory
regime has intervened
How Many Variables
, variables that had previously been ignored turn out to have predictive value when used in
combination with other variables. For example, one credit card issuer, that had never included
data on cash advances in its customer profitability models, discovered through data mining
What Must the Data Contain
At a minimum, the data must contain examples of all possible outcomes of interest. In
directed data mining, where the goal is to predict the value of a particular target
variable, it is crucial to have a model set comprised of preclassified data
STEP THREE: GET TO KNOW THE DATA

Examine Distributions
A good first step is to examine a histogram of each variable in the dataset and think
about what it is telling you.
Compare Values with Descriptions
Look at the values of each variable and compare them with the description given for
that variable in available documentation. This exercise often reveals that the
descriptions are inaccurate or incomplete
Validate Assumptions
Using simple cross-tabulation and visualization tools such as scatter plots, bar graphs,
and maps, validate assumptions about the data
STEP FOUR: CREATE A MODEL SET

The model set contains all the data that is used in the modeling process.
Some of the data in the model set is used to find patterns.
Some of the data in the model set is used to verify that the model is stable
Creating a model set requires assembling data from multiple sources to form customer signatures and then
preparing the data for analysis.
Including Multiple Timeframes
The primary goal of the methodology is creating stable models. Among other things, that means models that
will work at any time of year and well into the future. This is more likely to happen if the data in the model set
does not all come from one time of year
Creating a Model Set for Prediction
Although the model set should contain multiple timeframes, any one customer signature should have a gap in
time between the predictor variables and the target variable. Time can always be divided into three periods:
the past, present, and future. When making a prediction, a model uses data from the past to make predictions
about the future.
Partitioning the Model Set
Once the preclassified data has been obtained from the appropriate timeframes, the
methodology calls for dividing it into three parts.
The first part, the training set, is used to build the initial model.
The second part, the validation set, is used to adjust the initial model to make it more
general and less tied to the idiosyncrasies of the training set.
The third part, the test set, is used to gauge the likely effectiveness of the model when
applied to unseen data.
STEP FIVE: FIX PROBLEMS WITH THE DATA

Categorical Variables with Too Many Values

Numeric Variables with Skewed Distributions and Outliers
Missing Values
Values with Meanings That Change over Time
STEP SIX: TRANSFORM DATA TO BRING
INFORMATION TO THE SURFACE
Once the data has been assembled and major data problems fixed, the data must
still be prepared for analysis.
This involves adding derived fields to bring information to the surface.
It may also involve removing outliers, binning numeric variables, grouping classes for
categorical variables, applying transformations such as logarithms etc.
STEP SEVEN: BUILD MODELS

This explanation may take the form of a neural network, a decision tree, a linkage
graph, or some other representation of the relationship between the target and the
other fields in the database.
In undirected data mining, there is no target variable. The model finds relationships
between records and expresses them as association rules or by assigning them to
common clusters
STEP EIGHT: ASSESS MODELS

■■ How accurate is the model?

■■ How well does the model describe the observed data?
■■ How much confidence can be placed in the model’s predictions?
■■ How comprehensible is the model?
STEP NINE: DEPLOY MODELS
Deploying a model means moving it from the data mining environment to the scoring
environment.
To deploy the model, a programmer takes a printed description of the model and
recodes it in another programming language so it can be run on the scoring platform.

STEP TEN: ASSESS RESULTS

STEP ELEVEN: BEGIN AGAIN
DATA MINING THROUGH SOCIAL MEDIA
Social media analytics refers to the practice of gathering data from social media
platforms and analyzing the data to help decision makers address specific problems
Twitter is a real-time information network that connects users and followers to the
latest stories, ideas, opinions, and news. Likewise, Yelp is a site where customers
publish unsolicited reviews and viewers read them
Social media analytics can analyze social media data to obtain consumers’ innovative
ideas and enhance customer relationships.
EXAMPLES
From McDonald’s to Pepsi to Marriott, have employed social media analytics in order
to derive a competitive advantage
Coca-Cola Company operates a social media center called the Hub Network
(Journey Staff, 2014), which deploys a suite of social listening, analysis, and
engagement tools to deliver a better, faster, and more efficient response to real-time
social media opportunities and issues

Multinational hospitality company Marriott International operates its own social

media center, named M Live; here, about a dozen employees analyze Twitter feeds,
Instagram photos, and Facebook posts in effort to engage clients in social
conversations, increase the hotel chain’s brand presence, and keep up with the
latesttrends
SOCIAL MEDIA ANALYTICS
Social analytics is the key to ensure high-level performance of social media
campaigns, and also to make data-driven decisions. It helps social media team learn
what works and what does not work when it comes to content and strategy. This
analysis is based on the level of audience engagement, conversion, and outreach.
For example, the number of likes on post on Facebook defines the level of audience
engagement, whereas the number of clicks on a LinkedIn post indicates conversions.
Based on the likes and clicks, one can measure the performance of content and tweak
future content strategy. Similarly, when companiess analyze competitor’s social
networks, it gives an idea of what content is performing better.
SOCIAL MEDIA ANALYTICS
Social media analytics is an essential function of marketing that helps marketers
track, measure and analyze the performance of their social campaigns. It helps in
aligning the performance of social campaigns with marketing goals, and aides in
justifying the investments made in these campaigns.
TYPOLOGY OF SOCIAL MEDIA ANALYTICS
Real time customer Analytics
 Real-time customer analytics allow companies to manage conversations with customers,
respond to their issues, and deliver information that is timely and relevant via social media
channels in real time
 It is important to identify and resolve individualized incidents that may spread quickly via
word of mouth. Continuous monitoring of online channels is seen as a practice for preventing
personalized issues from damaging the company’s reputation. Improved customer intelligence
obtained through real-time customer analytics enables the company to formulate a more
precise, timely response
TYPOLOGY OF SOCIAL MEDIA ANALYTICS
Non-real-time customer analytics
 While real-time customer analytics is used to formulate reactive marketing
strategies, non-realtime customer analytics focuses on developing proactive
marketing strategies based on the analysis of customer data over time.
Non-real-time customer analytics may produce simple quantitative metric
Example
 frequency or duration of webpage visits and demographics and geolocation of
visitors, number and growth of members on a social media platform, and
conversion rate.
TYPOLOGY OF SOCIAL MEDIA ANALYTICS
 In addition to simple metrics, companies collect qualitative information about
sentiment or opinion.
 Using accumulated historical customer comments, companies can analyse how
customer behaviours evolve over time and may design marketing messages that
have higher reach and response
Data collection and processing can become unwieldy when a large number of social media
platforms are monitored.
For example, a trend analysis is more suitable for non-real time analysis, but consumers’
complaints may be monitored in real time.
The trend analysis of consumers’ responses to competitors’ products and services may provide
new or improved product ideas.
Companies can also investigate the relationship between consumer-generated contents and
sales performances.
Companies can access their best-performing competitor’s social media data. They can conduct
a trend analysis of customer sentiment, postings on social media platforms, number of
followers, demographics and geolocations of customers. The competitors’ metrics can serve as
a benchmark in developing marketing strategies and improving the home company’s products
and services.
Analyzing social media has become a popular research and business activity due to
the availability of user-generated content accessible through web-based APIs as
provided by major social media sites
Based on a survey of existing commercial tools, it is found find that sentiment analysis,
social network analysis, and statistical methods are most widely used.
SENTIMENT ANALYSIS
Sentiment analysis, also called opinion mining, refers to the application of computational
technologies such as natural language processing and computational linguistics to identify and
extract subjective information from vast amounts of user generated content
Two methods of Sentiment analysis
1. A machine-learning method with which the analytics learns to evaluate sentiment through
acquisition and integration of knowledge gained from a large number of sentiment examples
2. A lexical-based method with which the analytics evaluates sentiment by utilizing a large
dictionary of pre scored words and phrases.
Sentiment analysis has been used successfully for businesses activities, including predicting
stock market movements, determining market trends, analyzing product defects, and managing
crises
Sentiment context: to extract sentiment, one needs to know the context of the contents,
which may vary significantly from specialty sites to general sites where contents can
cover a wide spectrum of topics.
Sentiment level: text analytics can be conducted at the word, sentence, or document
level.
Sentiment orientation/polarity: the sentiment in a text can be positive, neutral, or
negative.
Sentiment strength: the strength of a sentiment in a text can be weak, moderate, or
strong.
Sentiment subjectivity: a given text may be an opinion or a fact
SOCIAL NETWORK ANALYSIS
Social network analysis is the process of analysing structures of social networks based
on social network theory, which seeks to explain how networks operate and analyse
the complex set of relationships within a network of individuals or organizations
PROCESS OF SOCIAL MEDIA ANALYTICS
Stage 1: Develop key social media metrics
 Extracting and delivering value from social media data requires precise measurement, so selecting the
right metrics is important in connecting data with desired outcomes (Grimes, 2013). Companies must
decide the objectives of social media analytics. Some of the widely-used objectives include building
brand awareness, increasing website visits, increasing revenue, enhancing corporate reputation, and
improving customer service/customer satisfaction.

These objectives must be accompanied by both quantitative metrics (e.g., number of

followers or fans, number and frequency of postings, number of likes or tweets) and
qualitative metrics (e.g., sentiment orientation, strength, subjectivity) necessary to
understand the attitudes and opinions of customers and competitors. Regarding
competitive analytics, the company must identify competitors to benchmark.
Stage 2: Choose, monitor, and listen to social media platforms
 Specific information obtained from engaging in social listening activities is highly dependent on where
we listen (e.g., selectively monitoring a single social media site vs. monitoring multiple sites) and how
the data is analyzed (Schweidel & Moe, 2012).
 Selection of social media platforms depends upon the impact of those platforms on the business (e.g.,
number of customers that use them) and accessibility of content on those social media platforms.
Forrester Research (2011) defines social listening as “programs that utilize social media to monitor,
measure, and respond to customer conversations and feedback online.”
Stage 3: Perform social media analytics

 The ability of decision makers to collect, filter, and interpret data, messages, and signals has a critical bearing on
their strategy (Makadok & Barney, 2001).
 Social media analytics aim to monitor, filter, and analyse the discussions that take place on social media platforms,
providing a comprehensive picture of consumer opinions regarding products and services.
 Sentiment analysis can suggest whether a product, service, or customer support is better or worse than industry
average.
 Correlating sentiment to recent changes in product design, for example, could provide essential feedback..
 While some social media data are structured, social media data are mainly unstructured and semistructured,
leading to high diversity, ambiguity, and textual disorder.
 These unstructured and semi structured data require pre-processing and scrubbing before data analytics to
eliminate missing data, incorrect data, and inconsistent data. T
 the process of data cleaning may involve spell checking, removal of typographical errors or duplicates, validating
and correcting values against a known list of entities, and tagging data with metadata.
Social media intelligence
 It describes what already has happened
 Social media intelligence enables managers to prescribe what should be done with the results of social
media analytics. Social media intelligence is achieved by combining knowledge generated from
traditional intelligence activities and knowledge gained from social media analytics, and helps
managers develop better actionable market decisions that align with the company’s objectives
 Social media intelligence enables managers to prescribe what should be done with the results of social
media analytics. Social media intelligence is achieved by combining knowledge generated from
traditional intelligence activities and knowledge gained from social media analytics, and helps
managers develop better actionable market decisions that align with the company’s objectives