Business Analytics, evolution & scope,

business analytics process, decision models

.Spreadsheets , spreadsheet functions,
spreadsheet modeling overview. Datasets ,
data cleaning, missing data handling
,outliers. Data visualization-charts and
graphs ,data queries-sorting and filtering
,summarizing data-frequency, relative
frequency ,histograms ,percentages and
quartiles, cross tabulation ,pivots.
 Everyone make decisions
 Managers in business makes numerous decisions everyday
 Decisions include what product to make & how to price
them, where to locate facilities, how many people to hire,
where to allocate advertising budgets, how to schedule
production etc…
 Many of these decisions have significant economic
consequences
 Managers need good information & assistance to make
critical decisions
 This is the reason why analytics is important in today's
business environment
 Business analytics simply “the use of data”
 Analytics “the scientific process/discipline of fact based
problem solving”
 Business Analytics “ the application of processes &
techniques that transform raw data in to meaningful
information to improve decision making”
 Business Analytics is the” combination of
skills,technologies,applications & processes used by
organisation to gain insight in to their business based on
data and statistics to drive business planning”
 Gaining insight in to business practices & customer
behaviors
 Improving predictability
 Identifying risk
 Improving effectiveness of communications
 Enhancing operational efficiency
 Taking right decisions at right time
 Faster &
Control cost better
decision
making
BA
Competitiv Understandi
e ng current
advantage trends
Knowing Customers & Managing Reputation

Improve Operational Efficiency

Financial Management

Proactive Risk Management

Manage Waste, Fraud &Abuse

 Knowing Customers & Managing Reputation

- For the Chief Marketing Officer or Director of Sales, understanding

customers can give one a competitive advantage
- Companies that understand why their customers are their customers can
fine tune their products/services, promotion, pricing policies, & store
locations and can enhance customer loyalty, target the most profitable
customer & improve customer satisfaction
- Companies collect data about the customers from social media
- Marketers collect information that people post publicity on sites such as
Pinterest, Facebook, Twitter etc..
- With this information company can detect negative opinions & take
proactive steps to address the problems & to build on positive opinions
 Improve Operational Efficiency
- Chief Operating Officers(COOs), responsible for daily operations of a
company , have traditionally been major users of analytics
- COOs must make sure their business operations are efficient & effective in
order to deliver

 Financial Management
- Data & analytics allow the Chief Financial Officer to gain more visibility &
control over the organization’s financial performance including revenue,
expenses, capital expenditures & profit/losses
- While monthly & quarterly processing of financials has been typical in the
past , companies are striving to automate their financial management
systems so that financial analysis can provide an up-to-date view of an
organisation’s financial posture
- Analytics can help the CFO track & evaluate the organisation’s
performance through the standard & dashboard displays
 Proactive Risk Management
- The goal of risk management is to allow organizations to make
risk-aware decisions that improve business performance
- Analytics can help to mitigate risks by bringing together data
from throughout the organisation & environment to identify
the risks

 Manage Waste, Fraud &Abuse

- Analytics can provide a minimum visibility in to the losses &
some analytics can be used to prevent the losses from occurring
 This is not a very new concept
 Modern evolution began with computers
 Computers facilitates the collection, management, analysis and reporting of
data Business Intelligence
 Business Intelligence dealt with
-How many units did you sell
-What products did customers buy
-How much money did they spend
-How many credit card transactions were completed
 BI helps create simple rules to flag exceptions automatically.
 Business Intelligence has involved into modern discipline called
Information Systems
 Modern Business Analytics stems from analysis and solution of complex
decision making problems using mathematical or computer based models
 Operations Research / Management Science
 OR/MS applications use modelling and optimization techniques for
translating real problems into mathematics, spread sheets or other
computer languages
 It helps identify best or optimal solution to a business problem
 Decision Support Systems combine BI concepts with OR/MS models to
create
analytical based computer systems to support decision making.
 DSS includes
-Data Management
-Model Management
-Communication Systems
 Descriptive analytics describes or summarizes a business’s
existing data to get a picture of what has happened in the
past or is happening currently.
 It is the simplest form of analytics and employs data
aggregation and mining techniques.
 This type of business analytics applies descriptive statistics
to existing data to make it more accessible to members
of an organization, from investors and shareholders to
marketing executives and sales managers.
 Descriptive analytics can help identify strengths and
weaknesses and provide insight into customer behavior.
 A look at past performance to determine what happened
and why
 Diagnostic analytical tools aid an analyst to dig deeper
into an issue at hand so that they can arrive at the source
of a problem
 It just provides an understanding of casual relationship
and sequences while looking backward
 Techniques that uses diagnostic analytics include
sensitivity analysis, regression analysis etc..
 Analysis of likely scenarios of what might happen
 Predictive analysis merely forecasts what are the
probabilities of the occurrence of the event
 It includes building and validation of models that provide
accurate predictions
 Techniques used
Data Mining
Simulation
 It can suggest all favorable outcomes according to a
specified course of action and also suggest various course
of actions to get a particular outcome
 This focus on providing decision-makers with support to
make decisions more effectively and to think more
constructively
 It can be obtained through the comparison of the
expected and achieved outcomes
 Techniques used
Decision Analysis
Simulation Optimisation
 Defining
the
business
needs
Updating
Explore
the
the data
system

Make a Analyse
decision the data

Predict
what’s
Optimise
likely to
happen
Step 1) Defining the Business Needs
 First stage in business analytics process
 It involves in understanding what the business would like to
improve on / the problem it wants to solve
 Key questions such as, "what data is available", "how can we
use it", "do we have sufficient data”

Step 2) Explore the Data

 This stage involves cleaning the data, making computations for
missing data, removing outliers , and transforming
combinations of variables to form new variable
 The removal of outlier is very important
 Once the data has cleaned, the analyst will try to make better
sense of data
Step 3) Analyse the Data
 Using statistical analysis methods analyst will find all
factors that are related to target variables
 On this stage data is cut, sliced and diced and different
comparisons are made while trying to derive actionable
insights from the data

Step 4) Predict What Likely to Happen

 Analyst will model the data using predictive techniques
 Then compare the predictive values with the actual values
and compute the predictive errors
Step 5) Optimize /Find the Best Solution
 At this stage the analyst will apply the predictive model
coefficients and outcomes to determine best solutions
with the given constraints and limitations
 The analyst recognize the model coefficients that are
most aligned to the organization’s strategic

Step 6) Make a Decision & Measure the Outcomes

 The analyst the make decisions and take action based on
the derived insights from the model and organizational
goal
 After a period of time the outcome action is measured
Step 7) Update the system with the results of the decision
 Finally the results of the decision and action and the new
insights derived from the model are recorded and
updated into the database.
 Information such as, ‘was the decision and action
effective?’, ‘how did the treatment group compare with
the control group?’ and ‘what was the return on
investment?’ are uploaded into the database.
 The result is an evolving database that is continuously
updated as soon as new insights and knowledge are
derived.
 A decision model is a logical or mathematical
representation of a problem or business situation that can
be used to understand, analyze, or facilitate decision
making.
 It describes the relationship between all the elements of
decisions
 This models are generally used to develop decision logic
/set of business rules
 It can be simple as directing relevant information to
appropriate decision makers at the right time
 It predicts what will happen if a given action is taken
 Decision modeling is the most advanced form of BA
 It predicts the outcomes of complex decisions for a
business
 The decision model first map the relationship between all
the elements of a decision
- Known data
- Decision one is considering
- Forecasted results of the decision
 SUM FUNCTION
=SUM(number1, number2,…)
 Average function
=Average(number2, number2,…)
 Count Function
=Count(Value1, value2,…)
 Max Function
=Max(number1, number2,…)
 Min function
=Min(mumber1,number2,…)
 SUMIF function
=SUMIF(RANGE,CRITERIA,SUM_RANGE)
 COUNTIF function
=COUNTIF(RANGE,CRITERIA)
 VLOOKUP function
 =VLOOKUP(lookup_value,table_array,col_index_num,[ra
nge_lookup])
 The most convenient and useful ways for business to
implement and analyze computer models
 An electronic spreadsheet is a block of occupied cells
 It is a file made of rows and columns that help sort data ,
arrange data easily and calculate numerical data
 It has an ability to calculate values using mathematical
formulas and the data in cells
 Examples ;
Microsoft excel,Lotus1-2-3 etc…
 Storing data

Purpose of Presenting
Planning
spreadsheet data

Gaining
insight
 You have probably used a spreadsheet to see what will
happen when you change a value.
 The sorts of things that you might have wanted to try out
could be:
 'If the cost of food rises by 10%, how much will that
decrease the profit we make for the school fete?"
 "If I reduce the number of staff in my business by two,
what effect will that have on the profit at the end of this
month?"
 Trying out different scenarios like this is called
'modelling'.
 Modelling works well in spreadsheets because all of the
calculations have been set up using formulas. So, when
you change one value e.g. food cost, all of the values
related to it, e.g. food cost, total costs, profit will change
automatically.
 A data set is a collection of related data that contains
individual data units organized(formatted) in a specific
way & may be accessed by one by one or more specific
access methods
 A data set is a collection of numbers or values that relate
to a particular subject. For example, the test scores of
each student in a particular class is a data set.
 The process which is used to determine inaccurate,
incomplete/unreasonable data and then improving the
quality through correction of detected errors and
ommissions.
 Highlight errors More efficient data
-removes major errors & -by cleaning up errors make
inconsistencies, blank cells, more efficient data
repetitions etc.

Benefits of data
cleaning

Better result More Accurate

- With efficient data we will be
able to quickly get what we
need from the data.
Missing Data
 This happened when there is no data value stored for a
variable in a observation
 This can be occur because of nonresponse
 Missing data is data missing for some (but not all)
variables and for some (but not all) cases
 If data are missing on a variable for all cases, then that
variable is said to be latent or unobserved
1) Missing Completely at Random

2) Missing at Random

3) Missing Not at Random

 Missing Completely at Random (MCAR)
- When we say data are missing completely at random, we
mean that the missingness has nothing to do with the
observation being studied (Completely Observed Variable
(X) and Partly Missing Variable (Y)).
- For example, a weighing scale that ran out of batteries, a
questionnaire might be lost in the post, or a blood sample
might be damaged in the lab.
- Generally, data are regarded as being MCAR when data
are missing by design, because of an equipment failure or
because the samples are lost in transit or technically
unsatisfactory.
 When we say data are missing at random, we mean
that missing data on a partly missing variable (Y) is
related to some other completely observed
variables(X) in the analysis model but not to the
values of Y itself.
 It is not specifically related to the missing
information. For example, if a child does not attend
an examination because the child is ill, this might be
predictable from other data we have about the child’s
health
 If the characters of the data do not meet those of
MCAR or MAR, then they fall into the category of
missing not at random (MNAR).
 When data are missing not at random, the
missingness is specifically related to what is missing,
e.g. a person does not attend a drug test because
the person took drugs the night before, a person did
not take English proficiency test due to his poor
English language skill.
A. Ignore the values
 If the size of the sample is large & deleting 1 or 2 sample
value doesn’t make any difference, the we can delete the
records
 Simply deleting/ignoring the variables

B. Fill the values in the average

 One of the simplest method
 Using average/mode value determined to fill in all the
missing values
 This method depends on the type of the problem
C. Fill the values with an attribute mean belonging to the
same bin

- If your data set has an attribute classified into categories

- Eg : income group as high, low & average and another
attribute,say,population has several missing values, we can
fill the missing values with the mean value of a particular
category
 Noisy data is data with a large amount of additional
meaningless information in it called noise.
 This includes data corruption and the term is often
used as a synonym for corrupt data.
 It also includes any data that a user system cannot
understand and interpret correctly.
 Outliers are data points having extremely large +ve or
–ve values, compared to the rest of data points
 It is a piece of data that is an abnormal distance from
other points
 There exist data objects that do not comply with the
general behavior/model of data. Such data objects, which
are completely different from or inconsistent with the
remaining set of data are called outliers
 Eg : in the following random numbers
1,99,100,101,103,109,110,201
1& 201 are outliers, 1 is the extreme low value & 201 is an
extreme high value
Point/Global outliers Contextual Outliers Collective Outliers

• If an individual data • If an individual data • If a collection of data

point can be points is anomalous
instance is with respect to the
considered anomalous in a entire data set, it is
anomalous with specific context or termed as a collective
respect to the rest condition (but not outlier.
of the data, then otherwise), then it
the datum is is termed as a
termed as a point contextual outlier
outlier.
 It refers to technologies that support visualization
 It includes digital images,graphs,videos,animations etc.
 Easy to implement
 Identify areas that need attention/improvement
Record
information Entering data

Explore
information Process & calculate

Explain Emphasize important aspects of

information
data
 Simple & easy to understand
 Charts are tools used to simplify reporting
 It clearly communicate the message
 Ability to compare data
a. Column chart

chart is organized in
vertical manner.

b. Bar chart-
chart is organized horizontally
(horizontal version of a column chart)
[Link] chart
- used to display trends
over time
-Shows the trends in 2 or more
datas

[Link] chart
Used to display the Proportional size of Various items
[Link](scatter)chart
-Chart is also known as
-Scatter chart Used to show the Relationship of Various
numeric values (between X&Y)

[Link] chart
It gives the amount of change occurred over a particular
time span
g. Doughnut chart
-Gives the relationship among the various parts
-storage capacity is more than pie chart

[Link] chart
-aggregate values of the various data series are compared
by the radar chart
[Link] chart
It is used to get the best
suited combination among
the two sets of data

[Link] chart
SORTING:
Method to arrange data in ascending or descending order.
FILERING:
Process to access specific data from the original one.
• While the regular data filter will filter the existing
dataset, you can use Excel advanced filter to extract
the data set to some other location as well.
• Excel Advanced Filter allows you to use complex
criteria. For example, if you have sales data, you can
filter data on a criterion where the sales rep is Bob
and the region is either North or South.
• You can use the Excel Advanced Filter to extract
unique records from your data.
Frequency
A frequency distribution is a summary table that shows the
number of occurrences of each value in a range

Relative frequency
The relative frequency of a class equals the fraction
/proportion of observations in each class

Relative frequency= frequency of class

n
Histograms
It is a graphical display of tabulated frequencies

Percentages
It is a fraction of 100 that is calculated by dividing the
numerator by the denominator and multiplying the result by
100

percentage = part
whole 100
Quartiles
-Quartiles in statistics are values that divide in to quarters
- The first quartile(Q1) is defined as the middle number
between the smallest number and the median of the data set
- The quartile (Q2)is the median of the data
-The quartile (Q3) is the middle value between the median
and highest value of the data set

PivotTabels
- It allows one to extract the significance from a large,detailes
data set
 It is the most useful analytical tools
 Also known as contingency table
 A cross tabulation is a 2 or more dimensional table that
records the number (frequency) of respondents that have
the specific characteristics described in the cells of the
table
 Cross tabulation is a statistical tool that is used to analyze
categorical data. Categorical data is data or variables that
are separated into different categories that are mutually
exclusive from one another.
1) Eliminates confusion while interpreting data

2) Helps in deriving innumerable insights

3) Offer data points to chart out a course of action

 Module 2 – Part 1
FINANCIAL ANALYTICS

1
Course Plan

Understanding Risk, credit risk analysis, fraud detection and prevention

analytics, analytics in banking and financial services - analytics in retail
banking and wealth management

2
Financial Analytics
• Financial analytics is a field that gives different views of a
company’s financial data.
• It helps to gain in depth knowledge and take action
against it to improve the performance of your business.
• Financial analytics has its effect on all parts of your
business.
• Financial analytics plays a very important role in
calculating the profit of a business.
• Financial analytics helps you to answer all your business
questions related to your business and also lets you to
forecast the future of your business.

3
Uses of Financial Analytics
• Understand the performance of an organization
• Measure and manage the value of tangible and intangible
assets of an organization
• Manage the investments of the company
• Forecast the variations in the market
• Increase the functionalities of information systems
• Improve the business processes and profits

4
Reasons for using Financial Analytics
1. Business Models
There are three new business models which form the
basis of financial analytics
i. Business to Business
ii. Business to Consumer
iii. Business to Employee
2. Changing role of the financial department
Most of the finance functions are automatic and requires
only fewer resources to manage them. This enables the
finance executives to concentrate more on the business
goals rather than just focusing on processing and
reconciling transactions.

5
Reasons for using Financial Analytics
3. Business Processes
Businesses are becoming more complex these days due to
the advancement of technologies. Lot of questions arise in
the mind of the business people. Analytics provide the
answers to all these questions. Financial analytics lets the
managers and executives in an organization to have access
to more accurate and detailed financial information of the
organization.
4. Integrated Analytics
These days companies use integrated financial analytics to
face the competition in the financial analytics market place.
Because of using such integrated financial analytics
companies will be able to analyze and share the information to
the sources inside and outside the organization. Organizations
should use integrated financial analytics to survive in the new
economy.
6
Risk Analysis
• A process that helps identify and assess potential threats
that could affect the success of a business or project.
• Includes means to measure, mitigate and control risks
effectively.
• An essential tool when the work involves threats and
risks.
• Many industries have recognized the increasing
importance of risk analysis such as:
 Medical
 Food and beverage
 Automotive
 Transportation
 Military
 Aerospace 7
Risk Analysis
• Many companies have established risk management
functions and procedures to perform risk analysis on a
continual basis.
• These companies may need to assess:
Their financial stability.
The financial feasibility of an investment.
The impact of new government policies.
The impact of new competitors coming into the market.
• Used to assess the potential health effects resulting from
human exposures to hazardous agents or situations.
• Widely used in manufacturing environments to improve
safety and manage potential risks in production lines.
• Used in all types of engineering of sophisticated systems to
ensure safety and reliability of systems, processes and 8
products.
Risk Analysis
• A key process area in project management.
• Helps deciding whether to proceed with a project or not.
• Ensures only projects with the highest chance of success
are selected.
• Used in project planning and during project
implementation to evaluate how a project can be
successfully completed.
• If risks are not considered and controlled, you will not be
able to minimize their impact on the schedule, scope, cost
or quality.
• It is possible for a project to be stopped for example if the
availability of resources become an issue, or the potential
benefits might not be sufficient.

9
Benefits of Risk Analysis
• Saves time and money.
• Reduces the level of uncertainty.
• Decreases the impact of negative events.
• Improves project controls.
• Improves organizational learning.

10
Applications of Risk Analytics
• Fraud detection and Prevention
• Market Liquidity risk Assessment
• Real-time Situational Awareness
• Product portfolio Analysis
• Credit Risk Analysis
• Benchmarking Risk Management

11
Stages of Risk Analysis
1. Risk identification.
2. Risk assessment.
3. Response planning and implementation.
4. Risk monitoring and control

12
Risk Identification
• Determining and documenting the potential risk that could
occur.
• An iterative process: As new risks may evolve or become
known as the project progresses.
• Identified risks and their characteristics are recorded in
the risk register.

13
Risk Identification Methods and Sources
• People who have gone through similar projects or events
• Expert opinions
• Failure history analysis
• SWOT analysis
• Assumption logs
• Observations and checklists
• Hazard analysis
• Scenario analysis
• Brainstorming.

14
Risk Assessment
• Helps to evaluate the significance of each risk.
• Highlight the risks that present the greatest threat on the
overall objectives.
• Risks should be prioritized according to their potential
impact and probability of occurring.
Risk Impact is the effect the risk will cause if it occurs.
Risk probability is a measure of the likelihood of the risk
occurring.

15
Risk Response Planning and Implementation
• You need now to respond to the assessed risks by
developing options and actions to reduce the probability
or impact.
• Here you will apply strategies to deal with them
effectively.
• This process should be:
Realistic
Cost effective
Agreed upon by key stakeholders
Owned by a responsible person.

16
Risk Response Planning and Implementation
Response Strategies:
1. Avoiding the risk
2. Transferring the risk
3. Mitigating the risk
4. Accepting the risk altogether

17
Risk Response Planning and Implementation

Avoiding the risk :

• It usually involves changing in the project plan such as:
Extending the schedule
Reducing the scope
Spending money or hiring resources to eliminate the risk.
• An example is when you hire a more skilled resource who
is likely to get the tasks done in less time.

18
Risk Response Planning and Implementation
Transferring the risk:
• Sharing the risk with someone else.
• It is simply handling off the risk to another team,
organization or a third party.
• Examples are:
Outsourcing a service
Buying an insurance

19
Risk Response Planning and Implementation
Mitigating the risk:
• Involves carrying out work now to reduce the probability
and/or impact of a risk to be within the acceptable
threshold limits.
• It may include preventive, detective or testing possible
ways to reduce the risk.
• Examples are:
Backing up the data to an offsite location
Choosing a more stable supplier

20
Risk Response Planning and Implementation

Accepting the risk altogether:

• An acceptable risk is the one that is tolerated because:
There is nothing you can do to prevent or mitigate it.
It is costly.
It is difficult to implement.
• One of the common acceptance strategies is to come up
with a contingency plan to cope with its consequences.

21
Controlling Risks
• Improves the efficiency of the risk analysis process.
• Involves:
Monitoring and re-assessing risks overtime.
Identifying new risks.
Evaluating the effectiveness of the risk response strategies.
• Performance information should be reviewed regularly:
Schedule progress
Costs incurred.
• Risks and risk response plans should be reviewed in regular
meetings to ensure plans are being implemented.
• In these meetings, key risks should be given more attention
and new risks should be raised and discussed

22
 Credit Risk
• The default risk on a debt that arises from a borrower who fails to
make the required payments is called Credit Risk.
• Any lender would include this as a first resort which includes
principal and interest along with disruption to cash flows and the
collection cost. The loss may be partial or even complete in many
cases.
• Higher borrowing costs are always associated with higher credit
risk levels in an efficient market.
• Due to this reason, the cost of borrowing can be used to conclude
credit risks based on the assessment by the participants of the
market.
• A credit check is performed by the lender to reduce this credit risk
on the prospective borrower and it may require the borrower to
take insurance which guarantees from a third party of the payment
to the lender.
• Credit risk increases when the borrowers, willingly or unwillingly,
are unable to pay. 23
Credit Risk Analysis
• The risks are calculated on the borrower’s ability to repay the
loan.
• To assess the risk credit risk the lenders, look at the five C’s of
the borrower.
• The five C’s are credit history, capacity to repay, capital, the
loans condition, and associated collateral.
• Some companies have a dedicated department only for
assessing the credit risk of its current and potential consumers.
• Due to the help of technology businesses can now analyze the
data quickly and assess customers risk profile.
• If an investor is evaluating to buy a bond, he will review the
credit rating of the bond before the purchase is made.
• If the rating is low then the issuer is considered to have a high
risk of default and alternatively, if it has a high rating then it is
considered to be a safe investment.
24
Types of Credit Risk
1) Credit Default Risk
• The risk of loss which arises from the debtor being
unlikely to repay the amount in full or when the debtor is
more than 90 days past is the due date of credit payment,
it gives rise to credit default risk.
• The Credit default risk impacts all the sensitive
transactions which are based on credit like loans,
derivatives or securities.
• Credit default risk is also checked by banks before
approving any credit cards or personal loan.

25
Types of Credit Risk
2) Concentration Risk
• This is the type of credit risk which is associated with
exposure of any single or group with the potential to
produce large losses to threaten the core operations of a
bank.
• It may arise in the single form of single name
concentration even industry concentration.

26
Types of Credit Risk
3) Country Risk
• The risk which arises from a sovereign state when it
freezes the payments for foreign currency overnight
defaults or its obligation which is termed as sovereign
risk.
• Country risk is exclusively associated with
the performance of macroeconomics of a country and is
also closely related to the political stability in the country.
• Sudden instability, which tends to happen during the
elections, results in high country risk.

27
Mitigation of Credit Risk
1. Risk-Based Pricing
• The lenders usually charge a higher rate of interest to
borrowers who are defaulters. This practice is known as
risk-based pricing.
• The lenders take into consideration the factors such as on
purpose credit rating and loan to value ratio.
2. Credit insurance and credit derivatives
• Bondholders hedge the risk by purchasing credit
derivatives or credit insurances.
• These contacts ensure the transference of the risk from
the gender to the server against a specific amount of
payment.
• Credit default swap is the most common form of credit
derivative used in the market. 28
Mitigation of Credit Risk
3. Covenants
• Stipulations may be written by lenders to the borrowers
which are called covenants.
• These are usually written into loan agreements such as a
periodic report about the financial condition, refrain from
paying dividends or further borrowing of amount or any
other specific action that affect the company’s financial
position in a negative way or repayment of the full loan at
the request of the gender in events such as borrower
changes or changes in debt to equity ratio or change in
interest coverage ratio.
4. Diversification
• Lenders diversify their borrower pools and reduce the
risk.

29
 Calculating Credit Risk
1. A standardized credit score such as FICO score is
determined of the borrower. The FICO score helps in
determining the credit history, repayment capacity and
creditworthiness of an individual. On one hand, the FICO
score indicates the way in which an individual makes the
repayment of his debts, it does not ensure repayment in the
future.
2. The next step in calculating credit risk would be to calculate
Debt-to-income ratio. This is calculated by monthly recurring
debts of a company and divided by gross monthly income.
The individuals who have a score of less than 35% are
considered as acceptable credit risks.
3. The last step is to factor in the potential loan of the borrower.
The potential loan would be the debt which can be taken by
the borrower on the basis of his credit cards and other
general creditworthiness. This gives a potential of loan and
payment capacity of the borrower. 30
Fraud
• Wrongful or criminal deception intended to result in
financial or personal gain.
• Fraud is an uncommon, well-considered, imperceptibly
concealed, time-evolving and often carefully organized
crime which appears in many types of forms.
 Credit card fraud
 Insurance fraud
 Corruption
 Counterfeit
 Product warranty fraud
 Healthcare fraud
 Telecommunications Fraud
 Money laundering
 Click fraud
 Identity theft
 Tax evasion
31
 Plagiarism
Fraud Triangle
• This basic conceptual model explains the factors that
together cause or explain the drivers for an individual to
commit occupational fraud.
• The model has three legs that together institute fraudulent
behavior:

32
Fraud Triangle
1. Pressure is the first leg and concerns the main motivation
for committing fraud. An individual will commit fraud
because a pressure or a problem is experienced of
financial, social, or any other nature, and it cannot be
resolved or relieved in an authorized manner.
2. Opportunity is the second leg of the model, and concerns
the precondition for an individual to be able to commit
fraud. Fraudulent activities can only be committed when
the opportunity exists for the individual to resolve or relieve
the experienced pressure or problem in an unauthorized
but concealed or hidden manner.
3. Rationalization is the psychological mechanism that
explains why fraudsters do not refrain from committing
fraud and think of their conduct as acceptable.

33
Fraud Detection and Prevention
• Fraud detection refers to the ability to recognize or
discover fraudulent activities, whereas fraud prevention
refers to measures that can be taken to avoid or reduce
fraud.
• The former is an ex post approach whereas the latter an
ex ante approach.

34
Benefits of Fraud Detection and Prevention
• Reduced exposure to fraudulent activities
• Reduced costs associated with fraud
• Find out the vulnerable employees at risk to fraud
• Have organizational controls
• Improves the results of the organization
• Gains the trust and confidence of the shareholders of the
organization

35
Methods of Fraud Detection
1. Sampling
2. Ad-Hoc
3. Repetitive or Continuous Analysis
4. Benford’s Law

36
Sampling
• Sampling is mandatory for certain processes of fraud
detection.
• Sampling will be more effective where there a lot of data
population involved.
• Disadvantage :
Sampling may not be able to fully control the fraud
detection as it takes only few population into
consideration.
Fraudulent transactions do not occur randomly therefore
an organization need to test all the transactions to
effectively detect fraud.

37
Ad-Hoc

• Ad-Hoc is nothing but finding out fraud by means of a

hypothesis.
• Test the transactions and find out if there are any
opportunities for fraud to take place.
• Have a hypothesis to test and find out if there is any
fraudulent activity occurring and then you can investigate
on the same.

38
Repetitive or Continuous Analysis
• Repetitive or Competitive Analysis means creating and
setting up scripts to run against big volume of data to
identify the frauds as they occur over a period of time.
• Run the script every day to go through all the transactions
and get periodic notification regarding the frauds.
• This method can help in improving the overall efficiency
and consistency of your fraud detection processes.

39
Benford’s Law
• Benford’s law can often be used as an indicator of
fraudulent data.
• Benford’s distribution is non-uniform with smaller digits
more likely than the larger digits.
• Using Benford’s law you can test certain points and
numbers and identify those which appear frequently than
they are supposed to and therefore they are the suspect.

40
Other Fraud Detection Methods
• Data Matching – This method will find out if there is any
data which exactly matches with another data.
• Sounds like – This is another powerful method where it
identifies variations of valid company employee names.
• Duplicates – This is another method which is most
commonly used by a lot of organizations to identify fraud
as well as any error occurring within all the business
transactions.
• Gaps – In this method you can find out the missing
sequential data.
For example if you have purchase orders which is issued
by the company in sequential order and if anything is
missing you can easily find out.

41
 Example – Fraud Detection
Three fraud detection methods used by Insurance company
1. Social Network Analysis (SNA) - SNA method follows the
hybrid approach to detect fraud. The hybrid approach includes
organizational business rules, statistical methods, pattern
analysis and network linkage analysis.
2. Fraud Detection Predictive Analytics for big data -
Predictive analytics uses text analytics and sentiment analysis
to look at big data for fraud detection. Predictive analysis has
been widely used by a lot of organizations as it helps in
proactively detecting frauds.
3. Social Customer Relationship Management (CRM) - Social
CRM is a process of fraud detection program. Linking social
media to CRM increases the transparency with the customers.
This transparency gains the customers trust over the
organization. This customer centric eco system benefits the
business to a great extent and also see through that the
customers are in control. 42
Steps of Fraud Detection
1. Perform SWOT
2. Build a dedicated fraud management team
3. Build or buy option
4. Clean data
5. Lay out relevant business rules
6. Setting the threshold
7. Predictive Modeling
8. Using SNA
9. Build an integrated case management system
leveraging social media
10. Forward looking analytics solutions

43
Steps of Fraud Detection
1. Perform SWOT:
• Many organizations have realized the increasing the
importance of fraud analytics.
• But in a hurry they are opting for expensive fraud detection
solutions that do not match with the company’s strengths and
weaknesses.
• Therefore organizations should do SWOT analysis before
starting with fraud detection program in order make it work to
the fullest.
2. Build a dedicated fraud management team:
• Traditional companies do not have a specific team for fraud
detection.
• But these days it is important to have a dedicated team who
works to find and prevent frauds in the organization.
• The team should have a proper flow and a proper reporting
fraud detection system. 44
Steps of Fraud Detection
3. Build or buy option:
• Once SWOT analysis is over and team allocation is done
it is important for the companies to decide how they want
to implement analytics and what resources are required.
• Companies need to know whether they are capable of
building an analytics solution for themselves or should
they purchase an analytical fraud detection solution from
an vendor.
• If there is a need to purchase then the company should
do a research about the different fraud detection vendors
and their products available in the market that fits their
company.
• There are few important factors to be considered while
purchasing fraud analytics solution like cost, user
interface, scalability, ease of integration and others. 45
Steps of Fraud Detection
4. Clean data:
• Integrate all the databases in the organization and
remove all unwanted things from the databases.
5. Lay out relevant business rules:
• Companies should come up with business rules after
doing a research on the resources and expertise of the
company.
• There are different types of fraud and few of which are
specific to particular industry.
• The external vendor cannot build a robust fraud detection
solution without getting the proper inputs from the
organization or company.

46
Steps of Fraud Detection
6. Setting the threshold:
• Whether the solution is in-built or purchased from outside the
company should provide boundary values for different anomalies.
• Thresholds are set using anomaly detection.
• If boundaries are set too high then there are chances of frauds to
slip through in between.
• If the boundaries are set too low then a lot of time and resources
are wasted.
• Therefore an organization should be very clever in determining
the thresholds
7. Predictive Modeling:
• Data mining tools are used to build models that produce fraud
propensity scores which is linked to unidentified metrics.
• After the scoring is done automatically, the results are established
for review and further analysis.
47
Steps of Fraud Detection
8. Using SNA:
• SNA has proved to be the most effective fraud detection program
by modeling relationships between various entities.
9. Build an integrated case management system leveraging
social media
• Case management system lets an investigator to know about all
the important findings that are relevant to an investigation and it
can be either structured or unstructured data.
• Metrics are the indicators of fraud and it can be helpful for
comparison at the organizational level or network level.
10. Forward looking analytics solutions
• Companies should always look out for any additional sources of
data and should integrate them with the current fraud detection
program to build the most efficient and effect fraud detection
program.
• This will help you to eradicate any new frauds that might develop
in the future. 48
Methods of Fraud Prevention
1. Neural Nets
2. Bayesian Models
3. Regression Analysis
4. Monitoring
5. Historical analysis “What-if”
6. Shadowing

49
Methods of Fraud Prevention
1. Neural Nets:
• Fraud detection methods based on neural network are
the most popular ones.
• The advantages of neural networks over other
techniques are that these models are able to learn from
the past and thus, improve results as time passes.
• neural networks are non-linear statistical data modeling
tools.
• They can be used to model complex relationships
between inputs and outputs or to find patterns in data.
• NNs can produce best result for only large transaction
dataset. And they need a long training dataset.

50
Methods of Fraud Prevention
2. Bayesian Models:
• Bayesian networks are directed cyclic graphs which
have nodes representing variables which may be
observable parameters or unknown variables.
• Each node is associated with a probability function that
takes as input a particular set of values for the node’s
parent variables and gives the probability of the
variable represented by the node.
• The Bayesian Network provides a compact
representation of joint probability distributions.
• The advantage is drawn from its representation power
and generalization.
• They also offer good generalization with limited training
data
51
Methods of Fraud Prevention
3. Regression Analysis:
• Regression analysis is a mathematical technique which
differs from other fraud detection techniques in that it
involves processing comparatively low volumes of
summary data through programs employing complex
logic.
• The mathematics involved in regression analysis can
be extremely complicated and the computations
involved, if performed manually, would be tedious.
• However, the underlying principles are easy for the
auditor to understand and the use of one of the many
statistical software packages now available will isolate
the auditor from the mathematics.

52
Fraud Analytics
• Fraud analytics is the combination of analytic technology
and Fraud.
• Analytics techniques with human interaction which will
help to detect the possible improper transactions like
fraud or bribery either before the transaction is done or
after the transaction is done.
• When analytics is added to traditional methods, it
enhances the fraud detection capabilities and gives a
new dimension to the fraud detection techniques.
• Fraud analytics also helps to measure the performance
which will help you to standardize and have a control for
constant improvement.

53
 Benefits of Fraud Analytics
• Identify Hidden Patterns - Fraud analytics identify new patterns,
trends and scenarios under which frauds take place, whereas
traditional approaches miss such things.
• Data Integration - It combines data from various sources and
public records that can be integrated into a model.
• Enhance existing efforts - Fraud analytics does not replace the
traditional rules based methods but it just adds up to your existing
efforts to bring you more improved results
• Harnessing unstructured data - Fraud analytics helps in deriving
the best value from unstructured data.
• Improve the performance - With the use of fraud analytics we
can easily identify what is working for your organization and what
is not working for your organization

54
Fraud Analytics Process Model

55
Fraud Analytics Process Model
• As a first step, a thorough definition of the business
problem is needed to be solved with analytics.
• Next, all source data must be identified that could be of
potential interest. This is a very important step, as data
are the key ingredient to any analytical exercise and the
selection of data will have a deterministic impact on the
analytical models that will be built in a subsequent step.
• All data will then be gathered in a staging area that could
be a data mart or data warehouse. Some basic
exploratory analysis can be considered here using for
instance OLAP(online analytical processing) facilities for
multidimensional data analysis (e.g., roll-up, drill down,
slicing and dicing).

56
Fraud Analytics Process Model
• This will be followed by a data cleaning step to get rid of
all inconsistencies, such as missing values and duplicate
data.
• Additional transformations may also be considered, such
as binning, alphanumeric to numeric coding, geographical
aggregation, and so on.
• In the analytics step, an analytical model will be estimated
on the preprocessed
• transformed data. In this stage, the actual fraud-detection
model is built.
• Finally, once the model has been built, it will be
interpreted and evaluated by the fraud Experts

57
Analytics in Banking & Financial Services

ASSIGNMENT

58
Analytics in Retail Banking
• Assignment

59
Module 3

2
Course Plan
Recruitment Analytics, On Boarding Analytics, Staffing
Analytics, Performance & Skill Gap Analytics,
Compensation & Benefit Analytics, Training & Learning
Analytics, Promotion and Succession Planning Analytics,
Compliance Analytics, Attrition & Retention Analytics

3
What is HR Analytics
• HR analytics is defined as the analytics of human resources
(employees), which embodies the entire life cycle of an
employee such as recruitment, managing performance,
incentives and employee engagement.

4
Purpose of HR Analytics
1. To identify the need for new departments and positions.
2. To determine which departments or positions can be
reassigned or eliminated.
3. To identify and quantify physical risks to employees in specific
positions.
4. To assign and delegate responsibility for tasks and goals.
5. To investigate the effectiveness of performance management
or performance-related pay in improving performance.
6. To evaluate the effectiveness of learning and development
activities.
7. To measure the impact of organizational development
interventions.
8. To improve employee retention rates.
9. To reduce absenteeism of employees.
10. To evaluate the effectiveness of different sources of recruits.
5
Levels of HR Analytics
3 levels of HR analytics
1. Descriptive Analytics:
• The use of data to record a particular aspect of HR and
provide information on what has been happening to. This
include:
a. Basic Workforce Data: Demographic data, working
arrangements, absence and sickness, turnover, health
and safety, pay.
b. People Development Data: learning programs,
performance assessment, skills and qualifications.
c. Perceptual Data: opinion surveys, focus group, exit
interviews.
d. Performance Data: financial, operational and customer
data's.
6
Levels of HR Analytics
2. Multi-dimensional Analytics:
• The combination of different sets of data to establish
any relationships between them.
• E.g. Regression Analysis
3. Predictive Analytics:
• The most advanced form of HR analytics is to use the
data to predict trends and therefore provide guidance
on the future HR strategy.
• E.g. Recruitment tools predict high performers, and
increasingly companies are able to predict which
employee is likely to leave.

7
Methods of HR Analytics
1. Clustering:
• This method helps to investigate hidden group patterns.
 E.g. which common people characteristics can predict better
sales performance?
 Which cluster of recruitment sources can predict better people
retention?
2. Driver Analysis:
• This method helps to understand hidden relationships
betweens events or people or business characteristics.
 E.g. what is the impact of poor engagement on client
satisfaction?
 What is the impact of sales training on business revenue
department?

8
Methods of HR Analytics
3. Risk Analytics:
• This method helps to understand probabilities or the
likelihood of occurring events.
• E.g. which high performers are at risk to leave in the next
24 months?
• Will the reduction of training investments increase the risk
of employee turnover?
4. Forecasting:
• This method will try to understand the future trend
lines based on historical patterns.
• E.g. what will be the employee turnover in the coming 3, 6
or 12 months?
• What will be the typical first 6-months time-to-productivity
trend line of a call center new comer without call center
9
experience?
Tools for HR Analytics
1. Optimize the interview process with - HireVue:

10
Tools for HR Analytics
2. Define clearly your benchmarks with – Bersin by
Deloitte:

11
Tools for HR Analytics
3. Increase retention of Staff learning initiatives with -
Skillport:
A variety of learning modalities to match individual
learning styles.
Analytics, reporting, measurement.
Extensive search and discovery.
Personalized learning recommendations.
Social and collaborative learning
Mobile access through Skillsoft Learning App.
Delivers a full complement of learning options and
administration features.

12
Tools for HR Analytics
4. Intelligently manage the talent with – Saba:
• Saba provides companies across the globe with a cloud-
based learning and talent management application
designed to drive employee engagement, elevate skills
and improve business productivity.
• Saba provides traditional talent management features
that include course building, learning management,
performance management, goal tracking, succession
planning, and recruiting.

13
Importance of HR Analytics
1. Improved Hiring Decisions
2. Good Training
3. Better Insights
4. Stable Retention
5. Tracking Company Profitability

14
Recruitment Analytics is tracking, evaluating, gathering
and analyzing employee and candidate statistics which
help in making better hiring decisions.
15
Importance of Recruitment Analytics
1. Finding the right candidate
2. Refining sourcing
3. Gathering performance data
4. Analyze overall recruitment experience
5. Help with profile tuning
6. Identify best candidate sources
7. Analyze market trends

16
Metrics Used for Recruitment Analytics
1. Qualified Candidates
2. Time to Fill
3. Offer Acceptance Rate
4. Hires to Goal
5. Quality of Hire
6. Cost per Hire
7. Diversity Hiring Metrics
8. Retention Rates

17
Metrics Used for Recruitment Analytics
1. Qualified Candidates:
A candidate who is determined to be a good fit for a job
after a phone interview can be labeled a qualified
candidate.
Qualified Candidates = Total Candidates Per Opening
Who Move Past the Phone Screen Stage
2. Time to Fill:
• Your time to fill metric simply adds up the total number of
days an open job goes unfilled, and your average time to
fill looks at all unfilled positions over a set time period.
Time to Fill = Total Number of Days Job Is Available
and Unfilled
Average Time to Fill = Total Number of Days of Open
Jobs / Total Number of Jobs Open
18
Metrics Used for Recruitment Analytics
3. Offer Acceptance Rate:
• This recruitment metric tracks just how many of the offers
you extend to candidates are accepted, and it usually
kicks in after the application and interview processes
have finished.
Offer Acceptance Rate (%) = ( Number of
Acceptances / Number of Offers ) x 100
(or)
Offer-to-Acceptance Ratio = Number of Offers /
Number of Acceptances
4. Hires to Goal:
• it’s the total hires you need within a set time period to
reach a predetermined hiring goal.
Hires to Goal = Total New Hires / Hiring Goal 19
Metrics Used for Recruitment Analytics
5. Quality to Hire:
• Your quality of hire results should tell you the value that
new hires bring to your firm.
Quality of Hire (%) = (Job Performance + Ramp-up
Time + Engagement + Cultural Fit) / N
(All scored out of 100, N = number of indicators)
Overall Quality of Hire (%) = [Avg. Quality of Hire score
+ (100 – Turnover Rate)] / 2
6. Cost Per Hire:
• The cost per hire recruitment metric adds up all the
expenses that go towards hiring a new employee, and
demonstrates the value of your recruitment methods.
Cost per Hire ($) = [Total External Costs] + [Total
Internal Costs] / Total Number of Hires
20
 Metrics Used for Recruitment Analytics
7. Diversity Hiring Metrics:
• Diversity metrics are important, and not just from a legal
standpoint.
• Equal opportunity regulations prohibit discrimination in
hiring based on race, color, religion, sex, pregnancy,
national origin, age, disability or genetic information.
8. Retention Rates:
• Retention rate tracks the total employees who stay with
your firm over a given time frame, out of the total number of
employees you had when that period began.
Retention Rate (%) = (Total Employees Still Employed at
End of Period / Total Employees at Start of Period) x 100

21
 ANALYTICS
On-boarding Analytics Can be defined as the technique
through which the new employee acquire knowledge,
skills etc to become effective organizational member.
22
4 C’s of On-Boarding Analytics

23
Metrics of On-Boarding Analytics
1. Yield Ratios
2. Orientation Costs
3. New Hire Survey
4. New Hire Separation
5. Speed to Performance
6. New Hire Engagement
7. Program Effectiveness

24
Metrics of On-Boarding Analytics
1. Yield Ratios:
• Yield ratios are usually used to measure how many
candidates were hired from a total number of
applications.
• Yield ratios show what percentage of candidates pass
from one stage of the hiring process to another.

25
Metrics of On-Boarding Analytics

2. Orientation Costs:
• Knowing the cost of orientation allows the company to
make sure their resources are being used in the best way
possible.
Average orientation cost per employee =
[(Time*Pay*Number)+Department]/Number
Time = Amount of time spend in orientation
Pay = Average hourly rate per employee
Number = Total number of employees in orientation
Department = Average HR department cost

26
Metrics of On-Boarding Analytics
3. New Hire Survey:
• In this process, a survey can be sent to employees at
different times during the on-boarding process to gauge
feedback.
• Below are 3 sample questions that can be asked.
• They could be formatted for a “yes or no” response or on
a likert scale.
1. The manager was well prepared for the arrival on the
first day at work.
2. The work done today is the role that was explained
during the recruitment process
3. The materials provided at orientation, easy to follow or
not? If not, what materials were not clear?

27
Metrics of On-Boarding Analytics
4. New Hire Separation:
• Employee separation happens with all staff at some time.
• Whether it is through resignation or termination, it is critical
• HR staff effectively follow their organizations employee
termination processes.
5. Speed to Performance:
• Managers should take care the following steps to get their new
employees up to speed
1. Make the experience personal
2. Expose them to company culture
3. Outline expectations
4. Provide feedback
5. Check in regularly

28
Metrics of On-Boarding Analytics
6. New Hire Engagement:
• On-boarding is all about getting new hires engaged.
• Satisfied employees will keep the company afloat, but
engaged employees will help the company grow.
7. Program Effectiveness:
• An effective on-boarding program goes way beyond the
initial classroom training.
• Running an extended program that integrates classroom
learning with informal, extended on the job training takes
effort, resources, and know-how.
• By utilizing these three measurements together one will
gain a lot of information about the company’s on-boarding
process.

29
Key steps to effective employee on-boarding
1. Have a clear on boarding process in place
• Preparation for on boarding a new employee needs to
start well before they walk through the door. Be sure to:
Help new employees prepare
Provide a functioning workstation
Arrange orientation
Organize the necessary paperwork

30
Key steps to effective employee on-boarding
2. Schedule the first day
• Help a new recruit to settle in quickly by making them feel
welcome and relaxed with these tips:
• Announce the new hire with a welcome email to the
team and/or company at the start of the day.
• Introduce the team to your new employee and guide
them around the workplace to meet key people
throughout the business.
• Show them around the workplace.

31
Key steps to effective employee on-boarding
3. Organize training and mentoring
• Revisit goals and responsibilities
• Schedule training
• Review mentoring opportunities
4. Help them settle into their role:
• Successfully on boarding a new employee is not an
overnight process
• Take the time to observe them and ask questions - do
they understand the business and their role? Facilitate any
additional training that may be needed.
• Don’t overlook recognition.
• Keep the employee motivated and engaged by
celebrating success.
32
Staffing Analytics

Analytics enable staffing firm to measure ROI, track the

effectiveness of marketing campaigns, and make
important business decisions.

33
Role of Predictive Analytics in Staffing
1. Talent Acquisition
2. Talent Pipeline planning
3. Job Response Optimization
4. Customer Acquisition

34
Performance Analytics enables you to track, aggregate,
and visualize key performance indicators over time,
rather than reporting on a point in time.

35
Metrics used for Performance Analytics
1. Indicators
2. Breakdowns
3. Scorecards
4. Dashboards
5. Widgets
6. Data collector

36
Metrics used for Performance Analytics
1. Indicators:
• Indicators define a performance measurement taken at
regular intervals of a business service, an activity, or
organizational behavior.
• These performance measurements result in a series
of indicator scores over time.
2. Breakdowns:
• Breakdowns enable you to group or filter indicator scores
for more detailed analysis, such as to show separate
scores for each assignment group.
• The values for each breakdown are called
breakdown elements.
• Breakdowns are automated, manual, or external,
depending on where these elements come from.
37
 Metrics used for Performance Analytics
3. Scorecards:
• Scorecards display data for a single indicator and enable you to
perform detailed analysis of the indicator data.
• Each indicator has an associated scorecard created automatically.
4. Dashboards:
• A dashboard can have multiple tabs and each tab can hold one or
more widgets.
5. Widgets:
• A Performance Analytics widget ties an indicator to a visualization,
such as a trend line, a set of columns, or a pie chart.
• Within the widget, you can filter or group indicator scores by
breakdowns.
6. Data Collector:
• Performance Analytics uses scheduled jobs to collect and clean
scores and snapshots, and enables you to manually set or import
38
scores.
Importance of Performance Analytics
1. Drive Performance
2. Establish a single version of truth
3. Realize fast time to value
4. Align the organization with company goals
5. Services quality

39
Skill Gap Analytics

A simple skill gap analytics consists of a list of skills

required along with a rating of the employee’s level
for each skill.

Skills needed – Skills possessed = Skills Gap

40
Steps to develop a Skill Gap Analytics
1. Identify Business Goals
2. Collect data
3. Interpret the data and make recommendations
4. Prepare a training plan to address identified
skill gaps in the current environment

41
Compensation Analytics

Compensation analytics focuses on optimizing the

cost of a workforce to drive bottom-line growth.

42
Compensation Analytics
• This branch of analytics helps organizations craft an
employer brand that effectively communicates a winning
employee value proposition.
• A value proposition that includes both financial and non-
financial rewards.
• In effect, compensation analytics enables HR leaders to
reward high-performers adequately, to boost workplace
morale, engagement, and retention.

43
Methods used in Compensation Analytics
1. Market Data Comparison
2. Labor Cost Analysis
3. People Count Analysis
4. Retention Analysis
5. High Performer Analysis
6. Sales Compensation Analysis
7. Geographic Pay Analysis

44
Analytics used to determine the extend to which the
goals of training programs have been achieved

45
Methods used for Training Analytics
1. Kirkpatrick Model:

46
Methods used for Training Analytics
2. CIRO Model:

47
Methods used for Training Analytics
3. Cost-Benefit Analysis:

Stag • Determining Training Cost

e1

Stag • Identify Potential Saving Results

e2

Stag • Compute Potential Savings

e3
• Conduct Cost & Savings Benefits
Stag Comparison
e4

48
Methods used for Training Analytics
4. ROI in Training:

49
Learning Analytics

Learning analytics studies the impact of

training on trainees.

50
Importance of Learning Analytics
1. Quality Assurance
2. Intervention
3. Identify At-risk Sub-groups
4. Enable for the development and
introduction of Adaptive Learning
5. Foster a Data-Driven Mindset
6. Increase Employee Motivation
7. Hurdles and Opportunities

51
Promotion Analytics

It is used to determine whether a given employee will

be successful if he is promoted to a higher level
position.

52
Skills Required for Promotion analytics

• Statistics
• Campaign Measurement
Technical
• Online Analytics
Skills
• Database Technology

• Promotional Strategy,
General Targeting and Segmentation
Promotion • Campaign Design and
knowledge Management
• Experience-Based Knowledge

53
Basis For Promotion
1. On merit basis:
Based on employee’s skill, knowledge, ability, efficiency
etc.
2. On seniority basis:
Based on the duration of service or experience at the
same post.
3. On merit-cum-seniority basis:
Consider both employee’s skills & experience.

54
Metrics Used For Promotion Analytics

55
Succession Planning Analytics

Predicting who is capable of succeeding an

important position in the organization.

56
Steps of Succession Planning

57
Importance of Succession Planning Analytics
1. Understand the risk associated with probable
successors identified for key high potential workers.
2. Analyze the turnover details of portable successors.
3. Track key high potential workers with and without
succession plans.
4. Evaluate effectiveness of the succession planning and
management process based on the number of plans
filled by successors identified for succession plans.

58
Monitoring whether the company is obeying the rules &
regulations outside the organization
59
Methods for Compliance Analytics
1. Rules based monitoring: To identify the known frauds
& compliance risks.
2. Anomaly detection: Recognize new fraud &
compliance risks.
3. Network analysis: To identify potentially illegal
activities inside the organization.
4. Text analysis: To sense written documents for
analysis.
5. Visual analysis: To summarize the results for the
stakeholders.

60
Attrition Analytics

Employees leaving an organization is called

attrition 61
Reasons For Employee Attrition
• Bad relationship with the employers.
• Bad relationships with coworkers.
• Less salary.
• Unfair working times.
• The job did not meet expectations.
• Feeling undervalued.
• Lack of proper coaching & feedback.
• Lack of growth opportunities.
• Bad company culture.

62
How To Minimize Employee Attrition
• Hire the right people.
• Offer competitive pay & benefits.
• Motivate & praise the employees.
• Allow flexible timings.
• Give opportunities to learn new skills & knowledge.

63
Retention Analytics

Retention refers to how many employees stick around

over a given period of time.
64
Steps For Employee Retention
• Step 1: Churn prediction: Create models to identify the
employees at risk of leaving.
• Step 2: Attrition root cause: Identify the reason for attrition
• Step 3: Resignation segments: Compare how the
resignation rate varies with respect to location, age
groups, function etc.

65
Marketing Analytics
Module 4 A
Course Plan
Basics of marketing analytics, marketing decisions
models, characteristics, types and benefits of marketing
decisions models, Segmentation using factors analysis
and cluster analysis, regression and choice based
segmentation, positioning - perceptual maps: developing
perceptual map, multi dimensional scaling

2
Marketing Analytics
• The practice of measuring, managing and analyzing the
marketing performance of a firm so that the ROI can be
optimized and increased is called Marketing Analytics.
• Marketing analytics portrays the customer insights and
trends.
• The following reasons induce the use of marketing
analytics:
Getting information related with new marketing trends
Identifying successful programs and evaluating their
reasons of success.
Analyzing trends over time
Completely analyzing the ROI of each program
Forecasting the outcomes

3
Components of Marketing analytics

People

Components
Input
of
& Steps
Marketing
Output
analytics

Tools &
Technology

4
3-Step Methodology for Marketing analytics

5
Importance of Marketing Analytics
1. Gaining a full view of customers across channels
2. Becoming more effective and proactive
3. Personalizing customer and market engagements
4. Visualizing success across enterprise
5. Trending data as a strategic asset

6
Guidelines for Marketing Analytics
1. Use a balanced assortment of analytic techniques.
2. Assess your analytic capabilities, and fill in the
gaps.
3. Act on what you learn

7
Guidelines for Marketing Analytics
1. Use a balanced assortment of analytic techniques:

8
 Guidelines for Marketing Analytics
2. Assess your analytic capabilities, and fill in the gaps.
• Assessing your current analytic capabilities is a good next step.
After all, it’s important to know where you stand along the
analytic spectrum, so you can identify where the gaps are and
start developing a strategy for filling them in.
• For example, a marketing organization may already be collecting
data from online and POS transactions, but what about all the
unstructured information from social media sources or call-center
logs?
• Such sources are a gold mine of information, and the technology
for converting unstructured data into actual insights that marketers
can use exists today.
• As such, a marketing organization may choose to plan and budget
for adding analytic capabilities that can fill that particular gap. Of
course, if you’re not quite sure where to start, well, that’s easy.
Start where your needs are greatest, and fill in the gaps over time9
as new needs arise.
Guidelines for Marketing Analytics
3. Act on what you learn
• There is absolutely no real value in all the information
marketing analytics can give you – unless you act on it.
• In a constant process of testing and learning, marketing
analytics enables you to improve your overall marketing
program performance by,
• for example:
Identifying channel deficiencies.
Adjusting strategies and tactics as needed.
Optimizing processes.
Gaining customer insight.
• Without the ability to test and evaluate the success of your
marketing programs, you would have no idea what was
working and what wasn’t, when or if things needed to
10
change, or how.
Marketing Decision Models
• A marketing decision model provides a way to visualize
the sequence of events that can occur following
alternative decisions in a logical framework, as well as the
outcomes associated with each possible pathway.
• Decision models can incorporate the probabilities of the
underlying states of nature in determining the distribution
of possible outcomes associated with a particular
decision.
• These probabilities are known to the decision-maker, but
are critically important.

11
Characteristics of Marketing Decision Models
1. Generalized set of processes
2. Purpose
3. Assumptions
4. Relationship between variables

12
 Types of Marketing Decision Models
1. On the basis of structural characteristics:
• Verbal Models:
In verbal models, the variables and their relationships are
stated in prose form. Such models may be mere
restatements of the main tenets of a theory.
• Graphical Models:
Graphical models are visual. They are used to isolate
variables and to suggest directions of relationships but are
not designed to provide numerical results.
• Mathematical Models:
Mathematical models explicitly specify the relationships
among variables, usually in equation form.
Where y = degree of preference
a0 = model parameters to be estimated statistically
13
Types of Marketing Decision Models
2. On the basis of Managerial Questions:
• Descriptive Decision Models:
These models address the question, “what will happen if
you do X?”
• Normative Decision Models:
These models address the question, “What is out best
course of action in a given situation?”

14
Benefits of Marketing Decision Models
1. Improved consistency in decisions
2. More decision options
3. Assessing the impact of variables
4. Facilitating group decision making
5. Updates mental models
6. Improves allocation of resources

15
Target Marketing: STP Approach
• A target market is a group of consumers or organizations
most likely to buy a company’s products or services.
• Because those buyers are likely to want or need a
company’s offerings, it makes the most sense for the
company to focus its marketing efforts on reaching them.
• Marketing to these buyers is the most effective and
efficient approach.
• The alternative - marketing to everyone - is inefficient and
expensive.

16
Steps in Target Marketing

17
Market Segmentation
• Facilitates Right Choice of Target Market
• Facilitates Effective Tapping of the Chosen Market
• Makes the Marketing Effort More Efficient and
Economic
• Helps Identify Less Satisfied Segments and
Concentrate on Them

18
Bases for Market Segmentation
Geographic
Segmentation Psychographi
Demographic Behavioral
c
Segmentation Segmentation
Segmentation
• Location
• Size • Age & life cycle • Lifestyles • Occasions –
• Population changes • Personality Regular or
density • Gender • Values Special
• Climate • Marital status • beliefs • Benefits
• Income • User Status –
• Social class not interested,
potential , first
• Family size
time , regular ,
• Occupation ex-user
• Educational • Quantity
level consumed –
• religion light, medium,
heavy
• Buyer
readiness
stage
• Loyalty status
– hard core, soft
core, split,
switchers 19
• Attitude
Segmentation Methods
1. Factor Analysis:
• Factor analysis is a marketing research technique that
analyses a large number of variables and reduces them
to a smaller number of key factors to better explain a
given marketing situation.
• Factor analysis is useful in benefit and psychographic
research segmentation.
• The central aim of factor analysis is orderly
simplification of several inter-related measures using
quite sophisticated mathematics to test for or confirm
generalizations.
• the most important statistic for these procedures is the
correlation coefficient.

20
Segmentation Methods
• There are two major types of factor analysis used in
market segmentation. They are:
i. R Factor Analysis:
It reduces the amount of data by finding similarities in
response to particular variables.
II. Q Factor Analysis:
It finds grouping of people that respond similarly to
research issues.

21
Segmentation Methods
2. Cluster Analysis:
• Cluster analysis is especially useful for market
segmentation.
• Segmenting a market means dividing its potential
consumers into separate sub-sets where
• Consumers in the same group are similar with respect to
a given set of characteristics
• Consumers belonging to different groups are dissimilar
with respect to the same set of characteristics
• This allows one to calibrate the marketing mix differently
according to the target consumer group.
• Help marketers discover distinct groups in their customer
bases, and then use this knowledge to develop targeted
marketing programs

22
Segmentation Methods
• The underlying definition of cluster analysis procedures
mimic the goals of market segmentation: -
to identify groups of respondents that minimizes
differences among members of the same group - highly
internally homogeneous groups
while maximizing differences between different groups -
highly externally heterogeneous groups
• Market Segmentation solution depends on variables used
to segment the market method used to arrive at a certain
segmentation

23
Importance of Market Segmentation
1. Adjustment of product and market appeals
2. Better position to spot marketing opportunities
3. Allocation of marketing budget
4. Understanding and meeting the needs of consumers
5. Stronger positioning
6. Enhanced efficiency
7. Competitive advantages
8. Targeted media
9. Market expansion
10. Better communication
11. Increases profitability
12. Identifies new market
13. Reduces cost
24
14. Reduces credit risks
Positioning
• Market Positioning refers to the ability to
influence consumer perception regarding a brand or
product relative to competitors.
• The objective of market positioning is to establish the
image or identity of a brand or product so that consumers
perceive it in a certain way.
• For example:
A handbag maker may position itself as a luxury status
symbol
A TV maker may position its TV as the most innovative
and cutting-edge
A fast-food restaurant chain may position itself as the
provider of cheap meals

25
Tasks involved in Positioning
Competitor’s Identification

Determining how Competitor’s are perceived and

evaluated

Determining Competitor’s position

Analyzing customer’s preferences

Making the positioning decision

Monitoring the position

26
Positioning Strategies
1. Attribute Positioning
2. Price/Quality Positioning
3. Use or Application Positioning
4. Product user Positioning
5. Usage and use time Positioning
6. Product class Positioning
7. Category Positioning
8. Benefit Positioning
9. Price-quality Positioning
10. Competitive Positioning
11. Corporate identity Positioning
12. Brand endorsement Positioning
27
Positioning using Multi-Dimensional Scaling
• Multidimensional Scaling (MDS) is a class of procedures
for representing perceptions and preferences of
respondents spatially by means of visual display.
• Perceived psychological relationships among stimuli are
represented as geometric relationships among points in
multidimensional space.
• These geometric representations are often called spacial
maps.
• Multidimensional scaling are use for:
To determine the number and nature of dimensions
consumers use to perceive different brands.
To position brands on these dimensions.
To identify the position of consumer’s ideal brand.

28
Objectives of MDS
1. Develop techniques of full text analysis
2. Provides visual presentation of similarities
3. Help explain observed similarities
4. Applied to any kind of distances

29
Applications of MDS in Marketing
• Brand image measurement
• Market segmentation
• New product development
• Assessing advertising activities and its effectiveness
• Distribution channel decisions

30
Perceptual Mapping/ Positioning Map
• Firms use perceptual or positioning maps to help them
develop a market positioning strategy for their product or
service.
• As the maps are based on the perception of the buyer
they are sometimes called perceptual maps.
• Positioning maps show where existing products and
services are positioned in the market so that the firm can
decide where they would like to place (position) their
product.
• Firms have two options they can either position their
product so that it fills a gap in the market or if they would
like to compete against their competitors they can
position it where existing products have placed their
product.

31
Perceptual Mapping/ Positioning Map
• The diagram below is a Perceptual Map of UK chocolate
confectionery Brands

32
Perceptual Mapping/ Positioning Map
• Perceptual maps can help identify where (in the market)
an organization could position a new brand.
• In our example this could be at the medium price and
medium quality position, as there is a gap there.
• There is also a gap in high price low quality but
consumers will not want to pay a lot of money for a low
quality product.
• Similarly the low price high quality box is empty because
manufacturers would find it difficult to make a high quality
chocolate for a cheap price or make a profit from selling a
high quality product at a low price.

33
Developing Perceptual Maps
1. Selecting determinant attributes
2. Listing the competitors
3. Market survey
4. Plotting the results
5. Checking for errors

34
Benefits of Perceptual Maps
1. Checks Reality
2. Impact of campaigns
3. Monitors new products
4. Monitors competition
5. Look for gaps
6. Understands segments
7. Track preference changes

35
Importance of Positioning
• Placing the product in customers mind
• Connects product offerings with target market
• Product cannot be everything to everyone
• Creates a locus in customers mind
• Providing competitive advantages
• Better serving and covering the market

36
Module 4 – Part 2

2
Course Plan
Click stream analytics, engagement quantification
frameworks, anonymous vs. registered users analysis,
Social Media Analytics - User generated content -
Sentiment Analysis- Analytics in digital decoding
consumer intent, decoding customer sentiments from
comments, Text mining from opinion platforms

3
 Web Analytics
• Web analytics is the measurement, collection, analysis and
reporting of web data for purposes of understanding and
optimizing web usage.
• Web analytics is not just a process for measuring web
traffic but can be used as a tool for business and market
research, and to assess and improve the effectiveness of
a website.
• Web analytics applications can also help companies
measure the results of traditional print or
broadcast advertising campaigns.
• It helps one to estimate how traffic to a website changes
after the launch of a new advertising campaign.
• Web analytics provides information about the number of
visitors to a website and the number of page views.
• It helps gauge traffic and popularity trends which is useful
4
for market research.
Web Analytics Process

5
Web Analytics Terminology
1. Hit:
A hit is a request to a web server for a file. There may
be many hits per page view since an HTML page can
contain multiple files, such as images.
2. Pageview:
A request for a file, or sometimes an event such as a
mouse click, that is defined as a page in the setup of
the web analytics tool.
3. Event:
A discrete action or class of actions that occurs on a
website.
4. Vist/session:
A visit or session is defined as a series of page
requests or, in the case of tags, image requests from
6
the same uniquely identified client.
Web Analytics Terminology
5. Bounce Rate:
The percentage of visits that are single page visits and
without any other interactions (clicks) on that page.
6. Click Path:
the chronological sequence of page views within a visit or
session.
7. Unique user:
The uniquely identified client that is generating page
views or hits within a defined time period (e.g. day, week
or month).
8. Active Time / Engagement Time:
Average amount of time that visitors spend actually
interacting with content on a web page, based on mouse
moves, clicks, hovers and scrolls. 7
Web Analytics Terminology
9. Average Page Depth / Page Views per Average
Session:
Page Depth is the approximate "size" of an average visit,
calculated by dividing total number of page views by total
number of visits.
10. Average Page View Duration:
Average amount of time that visitors spend on an average
page of the site.
11. Exit Rate / % Exit:
A statistic applied to an individual page, not a web site.
12. First Visit / First Session:
A visit from a uniquely identified client that has
theoretically not made any previous visits.

8
 Web Analytics Terminology
13. Frequency / Session per Unique:
Frequency measures how often visitors come to a website in a
given time period.
14. Impression:
an instance of an advertisement appearing on a viewed page.
15. New Visitor:
A visitor that has not made any previous visits.
16. Page Time Viewed / Page Visibility Time / Page View
Duration:
The time a single page is on the screen, measured as the
calculated difference between the time of the request for that
page and the time of the next recorded request. If there is no
next recorded request, then the viewing time of that instance of
that page is not included in reports. 9
 Web Analytics Terminology
17. Repeat Visitor:
A visitor that has made at least one previous visit.
18. Return Visitor:
A Unique visitor with activity consisting of a visit to a site
during a reporting period and where the Unique visitor
visited the site prior to the reporting period.
19. Session Duration / Visit Duration:
Average amount of time that visitors spend on the site each
time they visit. It is calculated as the sum total of the
duration of all the sessions divided by the total number of
sessions.

10
Web Analytics Terminology
20. Single Page Visit / Singleton:
A visit in which only a single page is viewed (this is not a
'bounce').
21. Site Overlay:
a report technique in which statistics (clicks) or hot spots
are superimposed, by physical location, on a visual
snapshot of the web page.

11
Data Collection Methods
• There are two major methods for collecting data for web
analytics:
1. Log Files Analysis:
• The older of the two methods, simply counts the hits
made in the web server logs and stores the data in an
easily-readable, easily-managable format.
• This method is based on server-side data collection;
there is nothing stored on the visitor’s computer,
nothing that runs in their browser.
2. Page Tagging:
• Concerns about the accuracy of log file analysis in the
presence of caching, and the desire to be able to
perform web analytics as an outsourced service, led to
the second data collection method, page tagging.
12
Data Collection Methods

13
 Data Collection Methods – Logfile Analysis
Advantages Disadvantages
• Does not require • Can only record
changes to the website interactions with the web
or extra hardware server.
installation. • Server must be
• Doesn't require extra configured to assign
bandwidth. cookies to visitors.
• Freedom to change • Only available to
tools with a relatively companies who run their
small amount of own web servers.
hassle. • Cannot log physical
• Logs both page locations.
request successes and
14
failures.
Data Collection Methods – Page Tagging
Advantages Disadvantages
• Near real-time • Requires extra code
reporting. added to the website.
• Easier to record • Uses extra bandwidth
additional information. each time the page
• Able to capture visitor loads.
interactions within flash • Can only record
animations. successful page loads,
not failures.
• Hard to switch analytic
tools.

15
Web Analytics
Advantages Disadvantages
• Traffic • Time Consuming
• Page Views • Problems due to
• Advertising certain types of
visitors.
• Optimize Website
• E-Marketing Plans

16
 Clickstream Analytics
• On a Web site, clickstream analysis (also called clickstream
analytics) is the process of collecting, analyzing and
reporting aggregate data about which pages a website
visitor visits and in what order.
• The path the visitor takes though a website is called the
clickstream.
• There are two levels of clickstream analysis, traffic analytics
and e-commerce analtyics.
• Traffic analytics operates at the server level and tracks how
many pages are served to the user, how long it takes each
page to load, how often the user hits the browser's back or
stop button and how much data is transmitted before the
user moves on.

17
 Clickstream Analytics
• E-commerce-based analysis uses clickstream data to determine
the effectiveness of the site as a channel-to-market. It's
concerned with what pages the shopper lingers on, what the
shopper puts in or takes out of a shopping cart, what items the
shopper purchases, whether or not the shopper belongs to
a loyalty program and uses a coupon code and the shopper's
preferred method of payment.
• Because an extremely large volume of data can be gathered
through clickstream analysis, many e-businesses rely on big
data analytics and related tools such as Hadoop to help interpret
the data and generate reports for specific areas of interest.
• Clickstream analysis is considered to be most effective when
used in conjunction with other, more traditional, market
evaluation resources.

18
Engagement Quantification Frameworks
• Engagement is the process by which government,
organizations, communities and individuals connect in the
development and implementation of decisions that affect
them.
• It is used as a tool to achieve outcomes, develop
understanding, educate and agree to solutions on issues
of concern.
• The level of engagement appropriate for each situation
can range from a one-way transfer of information through
to consulting and even actively involving or empowering
stakeholders in the decision making process.

19
Effective Engagement

Clear, Relevant Transparent

and Timely Decision Inclusiveness
Communication Making

Collaboration &
Integrity
Co-operation

20
Engagement through Product Life Cycle

• Concept Planning
Stage 1
• Options Analysis
Stage 2
• Preliminary Design
Stage 3
• Detailed Design
Stage 4
• Construction
Stage 5

21
Benefits of Engagement
• Project and problem definition
• Solution testing and value management
• Risk quantification
• Risk mitigation
• Credibility and reputation

22
 Anonymous Vs Registered User Analysis
Anonymous User Registered User
User of a website, program or other Citizens that visit the portal without
systems who has previously registered. having registered.
Users have an account on intranet and Users are not known to intranet.
have logged in.
Identified by username Tracked by their IP addresses
The average of reputation is 59% The average of reputation is 49%
49.4% of the contributed content by 85.2% of the contributed content by
registered users have been deleted anonymous users have been deleted
over time over time

Only registered can reply on an issue Allowing anonymous users to do this is

inviting trouble
Users in the registered group will It is not possible for anonymous users
automatically inherit the permissions
one assigned to the anonymous group

Registered users can participate in a They don’t have such permissions

23
poll
Social Media Analytics
• It is the systematic and scientific ways to consume the
vast amount of content created by Web-based social
media outlets, tools, and techniques for the betterment of
an organization’s competitiveness
• Fastest growing movement in analytics

24
Social Media Analytics
• HBR Analytic Services survey (HBR, 2010)
75% of the companies did not know where their
customers are talking about them
31% do not measure effectiveness of social media
only 23% are using social media analytics tools
7% are able to integrate social media into marketing
• Measuring the Social Media Impact
Descriptive analytics – simple counts/statistics
Social network analysis
Advanced analytics – predictive analytics, text mining

25
Steps in Social Media Analytics
1. Think of measurement as a guidance system, not a
rating system
2. Track the elusive sentiment
3. Continuously improve the accuracy of text analysis
4. Look at the ripple effect
5. Look beyond the brand
6. Identify your most powerful influencers
7. Look closely at the accuracy of your analytic tool
8. Incorporate social media intelligence into planning

26
Sentiment Analysis
• The practice of applying Natural Language Processing
and Text Analysis techniques to identify and extract
subjective information from a piece of text.
• The process of analyzing unstructured text to extract
relevant information and transforming it into useful
business intelligence
• It determines if an expression is positive, negative, or
neutral, and to what degree
• It is an emerging field that attempts to analyze and
measure human emotions and convert it into hard facts

27
 Sentiment Analysis
• Also called opinion mining since it includes identifying
attitudes, emotions, and opinions of a company’s product,
brand, or service
• It is text analytics that looks at the face value of the words to
give them meaning - Gives insight into the emotion behind
the words
• Helps businesses monitor news articles, online forums and
social networking sites for trends in opinions about their
products and services
• Businesses have realized that not all opinions are equally
important- Some opinions carry more weight than others
• A negative tweet by Lady Gaga will have a much greater
impact than a tweet by an ordinary person
• It is a tool to allow users to generate ‘influence scores’ to
identify people, blogs, forums etc. that are important 28
Types of Sentiment Analysis

Manual Keyboard
Processing Processing

Natural
Language
Processing
29
Levels of Sentiment Analysis

30
Methods for Sentiment Analysis
1. The Naive Bayes classifier:
• The Naive Bayes classifier uses the Bayes
Theorem, that for our problem says that the probability
of the label (positive or negative) for the given text is
equal to the probability of we find this text given
the label, times the probability a label occurs,
everything divided by the probability of we find this text:

• Since the text is composed of words, we can say:

31
 Methods for Sentiment Analysis
• We want to compare the probabilities of the labels and choose
the one with higher probability.
• Since the term P(word1, word2, word3…) is equal for
everything, we can remove it.
• Assuming that there is no dependence between words in the
text (which can cause some errors, because some words only
“work” together with others), we have:

• With a training set we can find every term of the equation, for
example:
P(label=positive) is the fraction of the training set that is
a positive text;
P(word1|label=negative) is the number of times
the word1 appears in a negative text divided by the number of
times the word1 appears in every text. 32
Methods for Sentiment Analysis
2. The SVM Classifier:
• This classifier works trying to create a line that divides the
dataset leaving the larger margin as possible between
points called support vectors.
• As per the figure below, the line A has a larger margin
than the line B, so the points divided by the line A have to
travel much more to cross the division, than if the data
was divided by B, so in this case we would choose the
line A.

33
Methods for Sentiment Analysis
3. Multi-Layer Perceptron:

34
Methods for Sentiment Analysis
4. Clustering Classifier:

35
Text Mining
• Text Mining is also known as Text Data Mining.
• The purpose is too unstructured information, extract
meaningful numeric indices from the text.
• Thus, make the information contained in the text
accessible to the various algorithms.
• Information can extract to derive summaries contained in
the documents.
• Hence, you can analyze words, clusters of words used in
documents.
• In the most general terms, text mining will “turn text into
numbers”.
• Such as predictive data mining projects, the application of
unsupervised learning methods.

36
Areas of Text Mining

37
Text Mining Process

38
Text Mining Applications

39
Business Value

40
Steps to deliver Business Value

41
Module 5

1
 Predictive Analytics
• Predictive analytics is the branch of the advanced
analytics which is used to make predictions about
unknown future events.
• Predictive analytics Uses many techniques from data
mining, statistics, modeling, machine learning, and
artificial intelligence to analyze current data to make
predictions about future.
• Predictive modeling is a process used in predictive
analytics to create a statistical model of future behavior.
• Predictive analytics is the area of data mining concerned
with forecasting probabilities and trends.

2
Predictive Analytics

3
Predictive Analytics Process
Define Project

Data Collection

Data Analysis

Statistics

Modeling

Deployment

Model Monitoring
4
Advantages of Predictive Analytics

Optimizing
Detecting
Marketing
Fraud
Campaigns

Improving
Reducing Risk
Operations

5
 Applications of Predictive Analytics

Banking &
Oil, Gas & Health
Financial
Utilities Insurance
Services

Governments
Retail & Public Manufacturing
Sector

6
 Predictive Models
• Using data to make decisions and to take actions using
models that are empirically derived and statistically valid.
• It is a form of data mining technology that works by
analyzing historical current data and generating a model
to help predict future outcomes.
• Predictive models are created whenever data is used to
train a predictive modeling technique.
• Data + predictive modeling technique = Predictive
Models

7
 Principles of Predictive Models

Definition & Support Interference

Principles Principles

1. Principle of Similarity 1. Lack of Knowledge

Principle
2. Principle of Extensibility
2. Lack of Concern Principle
3. Principle of Robustness
3. Lack of Definition Principle
4. Principle of Fault
Tolerance
4. Lack of Engineering
5. Principle of Ease of Principle
Control
5. Lack of Responsibility
6. Principle of Completeness Principle

8
 Types of Predictive Models

1. Logic Driven Predictive Models:

• Logic-Driven Models: based on experience, knowledge,
and logical relationships of variables and constants
connected to the desired business performance
outcome situation.
2. Data Driven Predictive Models:
• Use data collected from many sources to quantitatively
establish model relationships Influence Diagrams
visually show how various model elements relate to one
another

9
Logic Driven Predictive Models

10
 Logic Driven Predictive Models
1. Single-Period Purchase Decisions:
• One time purchasing decision (Examples: selling t-
shirts at a football game, newspapers, fresh bakery
products, fresh flowers)
• Seeks to balance the costs of inventory over stock and
under stock
2. Multiple Time Period Models:
• Fixed-Order Quantity Models
 Event triggered (Example: running out of stock)
• Fixed-Time Period Models
 Time triggered (Example: Monthly sales call by sales
representative)

11
 Data Driven Predictive Models
1. Retail Pricing Markdown Model:
• Discount from a retail or selling price, creating sale or
volume discount prices
• Discounted/Marked-down prices must still be above
costs or we lose money on each sale!
• Markdowns are typically expressed as a percent
• Sale Price = Retail Price × (1 - Markdown % as a
decimal)

12
 Data Driven Predictive Models
2. Modeling Relationships and Trends in Data:

Linear function, y = a + bx

Logarithmic Function, y = ln(x)

Polynomial Function, y = ax² + bx + c

Power Function, y = axp

Exponential Function, y = abx

13
 Advantages of Predictive Models
1. Higher Fault Tolerance & System Reliability
2. Better Load Balancing
3. Faster Error Diagnosis, Recovery and Error
Aversion
4. Deeper understanding of Business Objectives and
Relationships
5. Ability to Address and Answer Business Strategy
Decisions
6. Better & More Reliable Strategic Planning

14
 Models Involving Uncertainty - DSS
• Decision Support Systems are computer-based
information systems that provide interactive information
support to managers and business professionals during
the decision making process.
• DSSs use:
1. Analytical models,
2. Specialized databases,
3. A decision makers own insights and judgments,
4. An interactive, computer based modeling process to
support semi structured business decisions.

15
 Models Involving Uncertainty - DSS
Components of DSS:

16
 Models Involving Uncertainty - DSS
• A decision support system involves an interactive
analytical modeling process.
• There are four basic types of analytical modeling
activities are involved in using a DSS:
1. What-if analysis ,
2. Sensitivity analysis,
3. Goal-seeking analysis, and
4. Optimization analysis.

17
Models Involving Uncertainty - DSS

18
 Analytics in Telecom
• The rapid rise in the use of smart phones and other
connected mobile devices has triggered a spurt in the
volume of data flowing through the networks
of telecom operators.
• It is necessary that the operators process, store, and
extract insights from the available data.
• Big Data analytics can help them increase profitability by
helping optimize network usage and services, enhance
customer experience, and improve security.
• Research has shown that the potential for telecom
companies to benefit from Big Data analytics is
substantial.
• The potential of Big Data, however, poses a challenge:
how can a company utilize data to increase revenues
and profits across the value chain, spanning network 19
 Analytics in Telecom
• Big Data analytics, for instance, enables companies to
predict peak network usage so that they can take
measures to relieve congestion.
• It can also help identify customers who are most likely to
have problems paying bills as well as those about to
change operators, thus exacerbating churn.
• Operators are usually advised against taking the usual
top-down approach when it comes to Big Data analytics,
which marks out the problem to be solved and then
seeks out the data that may help resolve it.
• Instead, the operators should focus on the data itself,
using it to make correlations and connections.
• If done correctly, the data could reveal insights that could
form the basis of more streamlined operations. 20
 Analytics in Location based Intelligence
Marketing
• Location intelligence is a tool to analyze the spatial
components of business data.
• LI is capable of combining business data with spatial
data and running complex analytics to offer location-
enabled business intelligence.
• It can be used to derive meaningful insights, discover
relationships and identify trends.
• Technologies like the Internet of Things (IoT), autonomous
vehicles and sensors are capturing information that has
never been captured before, creating entirely new avenues
for geospatial data collection.
• As a result, location-based analytics and platforms that can
process and detect trends and provide intelligence are
becoming more popular. 21
 Analytics in Location based Intelligence
Marketing
• Just to give a perspective, by 2020 there will be 20.4
billion connected “things” in use, up from 8.4 billion in
2017.
• This huge jump can be attributed to the growing
popularity of smart connected devices or also be termed
as the Internet of Things.
• IoT is already creating a buzz and in days to come will
lead to creation of an overwhelming amount of data
about what we do, how we do it, and where it happens.
• Since everything happens somewhere, an enormous
amount of the data that companies collect has a spatial
component, collecting and analyzing this type of location
data can be intimidating because it involves new
methods, technology, and talent. 22
 Analytics in Consumer Packed Goods
1. Direct Consumer Relationships
2. Mobile & Location Based Services
3. Demand Prediction
4. Demand Driven Supply Chain Management
5. Idea-to-Product Acceleration
6. Store Clustering
7. Price Optimization
8. Shelf-Space Allocation

23
 Analytics in Utilities
1. Asset Management
2. Reduce fraud and leakage using intelligence from
smart devices
3. Improve quality of service with predictive modeling
of grid performance
4. Improve drilling
5. Improve operations

24
 Analytics in Health Care
1. Comparative Effectiveness
2. Disaster Planning
3. Patient Flow
4. Radio Frequency Identification (RFID)
5. Genetics

25
 Analytics in Online Retail
1. Optimizing Marketing Mix
2. Customized Product Recommendations
3. Effective Overall Supply Chain
4. Predictive Search
5. Recommendations & Promotions
6. Fraud Management
7. Business Intelligence

26