Indian School of Business

Business Analytics using Data Mining

Academic Year & Term: 2021-22 Term: 6 (Mohali)

Instructor: Prof. Vandith Pamuru Office Hours: By appointment – Microsoft Teams

Affiliation: ISB
Email: vandith_pamuru@isb.edu Academic Associate: Tarun Vashishat

Course Description
One of the Canadian former ice hockey players, Wayne Gretzky, said this once “I skate where the puck is going
to be, not where it has been.” As someone can guess, a major component of this course deals with predictive
analytics, i.e., how to convert a business objective to a prediction problem, what tools and techniques are around
for prediction, and how to evaluate a prediction model. In addition to the predictive analytics, the course focuses
on tools meant for discovering structure in the data (read extract insights from the data) that help businesses in
making decisions and driving strategies. The course also touches upon some very powerful visualization
techniques, again meant to discover insights from the data.
Analytics is being used in almost every areas, e.g., FMCG, apparel industry, finance sector, consulting,
education, healthcare, and what not. The widespread proliferation of IT influenced economic activity leaves
behind a rich trail of micro-level data. Yet, most organizations are data rich but information poor. Emerging
technologies such as RFID, weblogs, social networks, website usage tracking and vast amounts of online
information (such as product ratings and bid histories) have the potential to reveal a lot about consumer, supplier
and competitor preferences to those that have the ears (read data-mining capabilities) to listen. The questions that
can be addressed using analytics are plenty: How can mobile companies use their customer database to predict
customer churn or to personalize SMS messages for improving customer service? How can financial institutions
use past loan data to predict the chance of defaulting for a new loan applicant? How can Bollywood use data on
movies to predict the next box-office hit? How can charities use data from a campaign in one location to target
the right people in another location? And how can politicians use databases of supporters to segment and best
target each audience? So, the knowledge acquired in this course will benefit not only those who plan careers in
pure-play analytics but also those who plan to work in applied fields like targeted marketing, predictive modeling,
strategic consulting, risk management, etc. The bottom line is - the analytics is going to give students an edge in
whatever areas they are going to work for.
In the course, we work with real business problems and real data. Students will learn the types of questions
that data mining can answer and the appropriate data mining tools for answering different questions. The emphasis
is on understanding the concepts behind a wide set of data mining techniques and their relation to specific business
analytics situations, rather than on mastering the theoretical underpinnings of the techniques. The course will
offer a practical experience of converting business objectives to data mining problems. Around the middle of the
course, students will have an “Ideation” project that solely focuses on formulating a data mining problem after
identifying a business objective. The course ends with a term project that requires the students to work on a data

1
mining problem with a real dataset. Students will have individual assignments during the term to prepare
themselves for the project.
An important feature of this course is hands-on learning using state-of-the-art business analytics software and
an Excel data mining add-in. All required data mining algorithms (plus illustrative data sets) are provided in an
Excel add-in, XLMiner. In addition, we will introduce TIBCO Spotfire, an industry leading data visualization
tool. Students are not required to have any extraordinary technical ability in Statistics or other methodological
areas to benefit from the course. The material is developed and presented in an intuitive manner with the objective
of making the students smart consumers of this widely applicable technology, in the managerial context.
Wherever appropriate, real examples will be used to motivate the topic being covered. However, sufficient
comfort level in dealing with data is necessary in order to appreciate the values of different techniques by doing.
It will not be required to write codes, however, it is expected that students taking this course are comfortable
dealing with data and using new software.
Students are expected to attend all classes, as well as, come prepared for each class to make the best use of
the classroom time as the course content is delivered in an interactive fashion and each lecture significantly builds
upon the materials covered in previous lecture(s).

Course Objective and Key Takeaways from the Course

Students taking this course will appreciate and be able to identify the enormous opportunities that currently exist
in providing business analytics services based on data mining techniques.
Upon successful completion of the course, students should possess valuable practical analytical skills that will
equip them with a competitive edge in almost any contemporary workplace. In particular, the knowledge acquired
in this course will benefit those who plan careers in analytics, targeted marketing, predictive modeling and
strategic consulting. More formally, the course will provide participants with the following skills and knowledge:
• Be aware of the business analytics’ potential in today’s data rich environment
• Gain a practical understanding of the key data mining methods of classification, prediction, data reduction
and exploration
• Know how to decide when to use which technique
• Understand how to implement major techniques using software
• Become a smart and critical consumer of data mining techniques
• Gain the intellectual capital required to provide business analytics services

Learning Goals
In addition to the course objectives listed above, students should expect to develop the following by the end of
the course:

1. Effective Oral Communication

Each student shall be able to communicate verbally in an organized, clear, and persuasive manner, and be
a responsive listener.
Assessment: Project presentations

2. Critical and Integrative Thinking

2
 Each student shall be able to identify key issues in a business setting, develop a perspective that is
supported with relevant information and integrative thinking, to draw and assess conclusions.
Assessment: Individual Assignments, classroom discussions, ideation proposal, team presentations

3. Awareness and Working in Teams

Each student shall demonstrate an ability to work effectively in a team, exhibiting behavior that reflects
an understanding of the importance of individual roles and tasks, and the ability to manage conflict and
compromise, so that team goals are achieved.
Assessment: Team project report and team presentations

Recommended Textbook
Data Mining for Business Analytics: Concepts, Techniques, and Application with XLMiner by Galit
Shmueli, Nitin R. Patel and Peter C. Bruce.

Required Software
• Analytical Solver Data Mining (formerly known as XLMiner), an Excel add-in.
• Business Intelligence and visualization tool TIBCO Spotfire Professional.

Optional Software
• R with R-Studio (an open-source software; no assignment or classroom exercise will involve R programming;
this is optional; however, additional asynchronous material will be provided for those interested)
• Python with Jupyter notebooks (Also an open-source software. Code will be provided to replicate the same
examples discussed in class.)

Session-wise Schedule
Session
Topics Deliverables Recommended Readings
ID
• Book Chapters 1 and 2 –
• Business Analytics Introduction and Overview of the
Applications Data Mining Process
• Introduction to • Article “A Predictive Analytics
Supervised and Primer” by Thomas H.
S01 Unsupervised machine Davenport; HBR, Sep 02, 2014
• Article “Where predictive
learning techniques
analytics is having the biggest
• Explaining vs. Prediction impact” by Jacob LaRiviere,
• Data Partitioning in Justin Rao, Preston McAfee,
Supervised Learning Vijay K Narayanan, and Walter
Sun; HBR, May 25, 2016
• Please install the necessary software, at least, the trial version before next class.
Week 1

• You’ll need the software, in class, on Session 02.

• Revise linear regression before next class. Materials are provided on LMS.

3
 • Book Chapters 5.1 and 5.2 –
• Linear Regression for
Evaluating Predictive
Prediction
S02 Performance
• Prediction goals and
• Book Chapter 6 – Multiple
performance
Linear Regression

• Individual
Please submit by Monday
Monday Assignment 1 –
noon.
15%

• Profiling vs.
Classification • Book Chapter 10 – Logistic
• Logistic Regression for Regression
Classification • Book Chapters 5.3, 5.4, and 5.5
S03
• Classification goals and – Evaluating Predictive
performance Performance
• Ranking goals and
performance
• Book Chapter 7 – k-NN
• K-NN
• Book Chapter 8 – The Naïve
• Naïve Bayes
S04 Bayes Classifier
Week 2

• CART
• Book Chapter 9 – CART

Please read before submitting the

individual assignment.
• Case: “Screening for Chronic
• Group Project Kidney Disease,” Darden
Proposal – 10% Business Publishing, University
Please submit by Monday of Virginia (UV0871)
Monday • Individual
noon.
Assignment 2 – Please read before submitting the idea
20% for the group project.
• Article: “12 predictive analytics
screw-ups” by Robert L. Mitchell;
Computerworld, Jul 24, 2013
• Neural Networks
• Cluster Analysis
• Book Chapter 15 – Cluster
o Hierarchical
S05 Analysis
clustering
Week 3

•
o K-means
clustering

4
 • Article: Cluster Analysis for
• Case Study: Mall of Segmentation, Darden Business
America Publishing, University of Virginia
S06 • Cluster characterization Book Chapter 3- Data
Data Visualization Visualization
•
Meet the instructor to discuss the submitted proposal for the group project.
• Individual
Please submit by Monday Assignment 3 –
Monday
noon. 15%

• Mining for patterns using

Association Rules
o Personalized
• Book Chapter 14 – Association
recommender
Rules and Collaborative Filtering
systems using
• Article: “Amazon.com
S07 Collaborative
recommendations: item-to-item
Filtering
collaborative filtering” by Linden,
• Reinforcement Learning
G., Smith, B., and York, J.
• The future of Data
Science

Please submit before the • Group Project

session starts. Presentation in
class – 10%
S08 Individual exercise
in class – evaluating
Week 4

Please be present in class to

evaluate other teams’ other teams’ project
presentations. presentations – 05%
• Group Project
Week 5

Please submit by Monday

Monday Final Report –
noon.
20%

NOTE: Classes will not be recorded, and attendance is mandatory.

Asynchronous components:
• Profiling using Logistic regression (complementary materials to Session 3)
• Ensemble (Complementary materials to Session 5)
• Multi-class classification (Complementary materials to Session 5)
• Guest Lecture: “Analytics at Scale” – Abhishek Kumar, Machine Learning, Google Inc
• Tutorial on R and Python programming, counterpart to all the methods covered in Sessions 2 through 7 (Optional
component for students)
5
 NOTE: Classes will not be recorded and attendance is mandatory

Graded Components
Name of the
component
(For details Take- Group Softcopy / Mark
Submission Coding Weight
on each home or Assignm Hard copy release
Deadline Scheme
component in-class ent (Y/N) submission date
stay tuned to
LMS posts)
Within 10
Individual Week 2 –
Take-home N Soft copy 2N-b days after 15%
Assignment 1 Monday noon
submission

Individual Week 3 –
Take-home N -do- 2N-b -do- 20%
Assignment 2 Monday noon
Group
Week 3 –
Project Take-home Y -do- 3N-b -do- 10%
Monday noon
Proposal

Individual Week 4 –
Take-home N -do- 2N-b -do- 15%
Assignment 3 Monday noon

Group
Project Session 8 In-class Y -do- 3N-b -do- 10%
Presentation
Individual
exercise in
class -
Session 8 In-class N -do- 4N -do- 05%
evaluating
other teams’
presentations

Group
Week 5 –
Project Final Take-home Y -do- 3N-b -do- 20%
Monday noon
Report

Class All sessions In-class N NA NA -do- 05%

Engagement
(Keeping the
videos on,
participation in
in-class quizzes)
Team: 40%

6
Individual: 60%

Group Composition for Group Assignment

a) Size: To be decided and shared with the class once the final class size is known
b) Composition: No constraint
c) Can groups be formed across different sections? No
Deliverables Guidelines (used for grading)
---------------------------------------------------------------------------------------------------------------------------------------
Individual Assignments x 3 (20% x 1 + 15% x 2 = 50%)
You must work individually on your assignments. Report & data files (Excel, Spotfire, or pdf) should be uploaded
to the LMS before the deadline (no hard copy needs to be submitted). Details on the assignment will be made
available on the LMS as the course progresses.
Late assignments are subject to a penalty of 1% per hour of delay. Request for extension of deadline will not be
entertained unless it is a medical emergency or alike and a medical certificate is produced in support of the
emergency.
---------------------------------------------------------------------------------------------------------------------------------------
Group Project Ideation (10%)
This is a team project. Group size will be decided based on the final class size. In this ideation proposal, each
team is required to come up with a business idea that can be approached using data mining applied to a data set.
The details on the dataset will be provided as the course progresses. Detailed guidelines on preparing the
proposal will be made available on the LMS.

Late submission will NOT be allowed. This is a team assignment.

If multiple team members indicate insufficient contribution of a member, appropriate penalties will be applied
to that member based on team’s inputs.
---------------------------------------------------------------------------------------------------------------------------------------
Group Project Execution (10%+20% = 30%)
This is a team project. Group size will be decided based on the final class size. As part of the project, each team
is required to work on a dataset with certain business objective working through the data mining problem and
solution, to recommendations. Details on the project to work on will be made available on the LMS as the
course progresses.
The key delivery components of the project are as follows
1. In-class Project Presentation (10%)
Each team presents their project to the class on Session 8.
2. Final Project Report (20%)
The project report details the team project, from the business problem through the data mining problem
and solution, to recommendations. It should be of professional standard. Bullet-points type reports are
strongly discouraged.

Late submission will NOT be allowed. This is a team assignment.

If multiple team members indicate insufficient contribution of a member, appropriate penalties will be applied
to that member based on team’s inputs.
7
---------------------------------------------------------------------------------------------------------------------------------------
Individual exercise in class – evaluating other teams’ presentations (5%)
• Details on the framework to evaluate an analytics project will be provided in class.
• Each student is required to be present on Session 8 and evaluate other teams’ presentations, in addition to
presenting their own projects on the same day.
Class engagement and classroom etiquettes (5%)
• For everyone who are attending the classes online, you are expected to keep your videos on for the whole
duration of the class. This contributes to your class engagement scores. In addition, your participation in the
in-class quizzes will also contribute to the class engagement scores. Note that the performance on these
quizzes doesn’t count towards your grade, and only your participation does.
• Learning is an interactive process. ISB students are admitted partly based on the experiences they bring to
the learning community and what they can add to class discussions. Therefore, attendance is an important
aspect of studying at the ISB. Late arrival is disruptive to the learning environment. Students are expected to
be in the classroom before the session begins.
• Both in an out of the classroom discussions are strongly encouraged. You are expected to enhance the
overall learning environment of the class by coming prepared, asking questions and sharing issues related to
your experience.
• You are required to bring laptops to the class for active classroom participation. However, it should be used
only when the instructor permits to do so. No other electronics or communication devices are allowed to be
used inside the classroom.

ISB Attendance Policy

Learning is an interactive process. ISB students are admitted partly based on the experiences they bring to the
learning community and what they can add to class discussions. Therefore, attendance is an important aspect of
studying here. Absence is only appropriate in cases of extreme personal illness, injury, or close family
bereavement. Voluntary activities such as job interviews, business school competitions, travel plans, joyous
family occasions, etc. are not valid reasons for missing a class. Late arrival and early departure are disruptive to
the learning environment; you should log-in the class before the scheduled start time and stay till the conclusion
of class. However, if due to an extenuating situation a student is forced to miss a class session, the same should
be notified to the respective Academic Associate with a copy to the Faculty and the ASA office along with
supporting documentary proof.
The ISB expects students to attend all class sessions in every course and watch/participate in all asynchronous
activities. Attendance cannot be linked to watching or participating in asynchronous activities. Attendance
during synchronous / in class sessions will be recorded However, if due to completely unavoidable reasons a
student is forced to miss synchronous/in-class sessions, the School policy is below:
• If a student misses up to 20% of sessions (synchronous + in-class) in a course, i.e. one session, there
will be no grade penalty.
• If a student misses more than 20% and up to 30% of sessions (synchronous + in-class) in a course, i.e.
two sessions, s/he will obtain a letter grade lower than that awarded by the faculty according to the
curve for the course.
• If a student misses more than 30% and up to 40% of sessions (synchronous + in-class) in a course, i.e.
three sessions, the student will receive a letter grade that is two levels lower than that awarded by the
faculty according to the curve for the course.
8
 • If a student misses more than 40% of sessions (synchronous + in-class) in a course, i.e. four or more
sessions, the student will receive an ‘F’ grade in that course.
• Important Note: Attendance is mandatory for Session 8.

Appendix: Coding scheme

What kinds of collaborative activities are

What material can be referred to?1
allowed?
Reference Can I discuss Can I discuss Can I refer to Can I refer to the case-study
s/Coding general concepts specific issues external solutions or problem set
Scheme and ideas relevant associated with the material?2 solutions?
to the assignment assignment with
with others? others?
4N N N N N
3N-a Y N N N
3N-b N N Y N
2N-a Y Y N N
2N-b Y N Y N
2N-c N N Y Y
1N Y Y Y N
0N Y Y Y Y

1
Any referencing needs to be accompanied with appropriate citations
2
A non-exhaustive list includes journal articles, news items, databases, industry reports, open courseware
9