Session 1 - Introduction To Data Analytics

TRAINING PROGRAM
ON DATA ANALYTICS
WITH POWER BI
Session 1
Delivered by
Dr. Pratyush Banerjee,
Associate Professor,
IMI Bhubaneswar
https://milvest.sambhavkadam.org www.sambhavkadam.org/courses-2
SESSION
OBJECTIVES
1. Introduction of Faculty
2. 2. Steps for Creating an
Analytics Culture
3. Concept Discussion:
Descriptive and Predictive
Analytics
4. use cases and challenges
Some facts about me:
• Present Designation- Associate Professor–IMI

Bhubaneswar
• Earlier associated with TAPMI, XIM Universitry,
Management Department, BITS-Pilani, Pilani Campus and
IBS Hyderabad, IFHE University,
Course Instructor: • Was associated with RPO and Telecom industries
Dr. Pratyush • Teaching HR Analytics since 2015, involved in curriculum
Banerjee development
[PhD • Certified HR Analytics Professional from Aon-Hewitt
(Management), Learning Center, 2017
PGDM (HR), B. Tech • Certified Business Analytics Professional from Carlton
(Electronics & Advanced Management Institute (UK), 2015
Telecomm.]
• Conducted workshops on HR and Business Analytics with
both Industry and Academia
• Written a book on HR Analytics titled “Practical
Applications of HR Analytics: A step by step Guide” by
Sage Publishers in 2019.
Defining Business Analytics
❑ Business Analytics may be defined as the application of

analytic logic to every facet of business function.
❑ Business Analysts use statistical applications, dashboards,
and optimization tools to analyze seemingly random and
unrelated data into meaningful and significant
relationships and then develop solutions for future planning
Popularized by the rise in capacity to store
data (cloud), advancement in computing
and competence of analytics as a
knowledge domain
Most major fortune 500 organizations are

using some level of analytics in their
The Rise of operations
Business
Analytics Major firms acquiring one or other analytics
suites provided by major players such as
Kenexa (IBM), Cornerstone (Xerox), Visier
(P&G) and Taleo (Oracle)
HOW MUCH DATA GETS PRODUCED
EVERY DAY?
10 Big Data and Analytics-A Few Examples
 Caesars Entertainment Corp. uses the big data on health-insurance claim of its
65,000 employees and their covered family members.
 Hardcastle, the franchise of McDonalds India, manages over 250 outlets across
India which collectively catered to approximately 650,000 customers each day.
 The HR team implemented complex decision algorithms to predict the
prospective ‘stayers’ vis-à-vis the potential ‘leavers’ based on the customer
data. This resulted in rise in Customer Satisfaction scores across majority of the
outlets from 30% to 40% in a year’s time.
Framing of queries or questions
Understanding appropriate data and metrics
Building appropriate platform for data Analytics

Steps for
creating
An Enhancement of Analytics Analytics capabilities
Analytics
Culture Disseminating value of Analytics
Being Data Driven – Case of P & G
P&G is among the foremost companies in the world in the use of data
and analytics. It is also a striking example of the impact of strong
leadership on establishing a data-driven culture in an organization.
In 2004, Filippo Passerini took over as CIO of P&G in 2004. He renamed

the IT department to “Information and Decision Solutions (IDS)”.
The renaming was based on Passerini’s belief that data and analytics
needed to play a more central role in decision-making at P&G.
Since then, the IDS unit has spearheaded several initiatives that have
transformed the way in which decisions are taken at P&G.
Initiatives taken at P&G to create a
data driven culture
Supporting Real-Time Decision-Making through “Decision Cockpits”:

 Passerini’s team developed “Decision Cockpits” – an initiative to provide a
single source of truth for data to all decision-makers across geographies
and business units.
 Decision Cockpits are dashboards that provide executives with visual
displays of data on business performance and market trends. The
dashboards can be customized according to individual needs.
 They allow executives to drill-down to granular views of data at a country,
brand or product-level and also provide real-time automated information
alerts.
 Decision Cockpits have been widely adopted at P&G with more than
58,000 executives using them every week. This in turn has helped P&G
speed up decision making and reduce time to market.
Analytics Cockpits at P&G
How Netflix uses data analytics to grow
its business?
 Netflix has digitized its interactions with its 151 million subscribers. It collects
data from each of its users and with the help of data analytics understands the
behavior of subscribers and their watching patterns.
 It then leverages that information to recommend movies and TV shows
customized as per the subscriber’s choice and preferences.
 As per Netflix, around 80% of the viewer’s activity is triggered by personalized
algorithmic recommendations.
 Where Netflix gains an edge over its peers is that by collecting different data
points, it creates detailed profiles of its subscribers which helps them engage
with them better.
What data does NETFLIX capture?
 Netflix collects information on how a user interacted and

responded to a TV show or a movie. If we go into details, it collects
the following data: –
• Time and date when a user watched a show
• The device used to watch the show
• If the user pauses the show, do they resume watching
• Does the user binge-watch an entire season of a TV show?
• If they do, how much time does it take to binge watch it?
What is the outcome of this analysis?
 Netflix has ratings that the viewer gives to the content they watch,
the number of searches they do, and what they search.
 The information collected is enough for creating a detailed profile
of a user, and this is exactly what Netflix does.
 It leverages data analytics to make a robust recommendation
algorithm that suggests the best content to the subscriber as per
their needs and preferences.
 The user no more must endlessly search through streams of
content to find out what he or she wants to watch.
 Netflix makes the job easier for them in the process, giving them a
better and customized viewer experience.
THE FINAL OUTCOME?
 The recommendation system of Netflix contributes to more than
80% of the content streamed by its subscribers (In contrast,
Amazon Prime’s retention rate is 75%, and Hulu’s is 64%) which has
helped Netflix earn a whopping one billion via customer retention.
 Due to this reason, Netflix doesn’t have to invest too much on
advertising and marketing their shows. They precisely know an
estimate of the people who would be interested in watching a
show.
 Netflix has emerged as the world’s most highly valued company,
with a total valuation of over $160 billion. Netflix can continue to
increase this valuation. It leverages its data by producing original
media and recommending the ideal content to viewers every time
they access the streaming platform.
Analytics in action at Google
 Fun workplace a data driven decision - major factor in attraction, retention, and
22
collaboration.
 Uses subtle tricks to encourage employees to eat healthy in terms control health care
costs
 Hide candy jars , move the salad counter to the front , place smaller plates
 Don’t just hire engineers - data indicated top performers and innovators now fluidly
move between industries
 Recognize great managers are essential for top performance and retention (Project
Oxygen)
 One on one coaching
 Frequent personalized feedback
 Retention algorithm and predictive modelling – take action before its too late
 Hiring algorithm – which candidate is going to be a good performer
 little value was added beyond four interviews, dramatically shortening time to hire
 Revisiting resumes – 1.5 % miss rate
 Use data to change preset opinions and to influence high performers
Analytics at Allrecipes
 Allrecipes is the world’s largest digital food brand, serving a global

community of 85 million home cooks across 18 web sites in 23
countries and 12 languages.
 This company uses data visualization to understand every stage of the
customer journey with the help of a software called Tableau.
 From analyzing web analytics to tracking content engagement, visual
analysis allows them to spot and stay on top of emerging trends.
 Visibility into customer and web analytics helps Allrecipes reach
dominant audiences like millennials, establishing a competitive
advantage in the digital landscape.
 The team also shares audience-specific insights with media and
advertising partners to support advertising partnerships while
maintaining a positive brand experience for the Allrecipes community.
Evolution of
Business
Analytics
Descriptive Analytics
Descriptive analytics: It encompasses the set of

techniques that describes what has happened in
the past.
Examples - data queries, reports, descriptive

statistics, data visualization (data dashboards),
and basic what-if spreadsheet models.
Comparing Proprietary Business
Intelligence Platforms in 2019
A Dashboard
created in MS
Power BI
A Social Media Engagement Dashboard

Predictive Analytics
Predictive analytics: It consists of techniques that use models constructed
from past data to predict the future or ascertain the impact of one
variable on another.
Statistical Applications- Cross-tabulation, frequency distribution, means

and SDs, Correlation, Regression, ANOVA/ t test, Chi square test
Machine learning Applications:

A. Classification - Unsupervised learning (Cluster analysis, Association,
Market Basket Analysis, self organizing Maps)
C. Prediction- Supervised Learning (Logistic Regression, Exponential
smoothing, Neural Networks, Naïve Bayes, SVM, Random Forests)
Key Applications of Predictive Analytics in Business
Business Objective Analysis Predictive Analytics
Techniques
Performance Group employees by performance and identify Cluster Analysis

Benchmarking differentiating KPIs that can be modified to incentivize
better performance
Root cause identification Identify factors driving sales, attrition analysis Cluster analysis,
classification
Business planning Identify impact of strategic alternatives Correlation and

Regression analysis
Predicting future outcomes Automated CV screening, joining prediction algorithms Association rule, Text
analytics, Classification
Employee / Customer Analyze large scale employee survey to understand Text analytics,
survey strategic policy related deficiencies Regression analysis
Training & Development Employee onboarding, personalized training A/B Tests, Text analytics,
Association rules
Industry Case study : Deloitte
• In 2015, Deloitte was the first big name to announce it was scrapping once-a-
year performance reviews, 360-degree feedback and objective cascading.
• This change occurred after the company calculated these processes were con-
suming a remarkable two million hours a year across the organization
(Buckingham and Goodall, 205, HBR)
• Deloitte’s new performance management process requires every team leader

to check in with each team member once a week to discuss near-term SMART
goals and priorities, comment on recent work and provide coaching.
HOW DELOITTE IS USING DATA TO AID PERFORMANCE DECISIONS
Source: Buckingham, M. and Goodall, A. (2015). Reinventing Performance Management, April, HBR
HOW DELOITTE IS USING DATA TO AID PERFORMANCE DECISIONS
Source: Buckingham, M. and Goodall, A. (2015). Reinventing Performance Management, April, HBR
How Lowe’s
uses data
driven
analytics
36 The Geeks Arrive In HR-Case of a

Canadian Bank
 A large Canadian bank suffering from theft and embezzlement in its branches
spent many years investing in training and monitoring tools to reduce fraud.
 Despite these ongoing programs, theft continued—and seemed particularly high
in smaller branches.
 The operations team, partnering with HR, embarked on a talent analytics project
to correlate patterns of loss against such factors as employee tenure, age,
experience, training, educational background, management demographics, and
geography.
 After many months of effort, the company found that the factor most correlated
to theft was the number of miles from the branch office to the district manager.
 People in this particular role who felt unsupervised were more likely to act
unethically.
 The bank reorganized its district managers to bring them closer to the branches,
and the loss rate dropped dramatically.
TECHNIQUES USEFUL FOR L&D ANALYTICS
Statistical tests – T Tests and AI & ML Based tests – Logistic

Analysis of Variance (ANOVA) regression and Decision trees
A decision tree output
Comparing licensed Data Analytics
Tools as of 2019
41 So, is Analytics the one stop solution to
all our problems?
 Data still requires human interpretation and expertise to
make such an impact
 If we rely on algorithms blindly, then there are chances
of the analysis backfiring on us.
 We need to be also mindful of using data ethically
 The focus should be on being cautiously optimistic about
the impact of analytics on business outcomes
Issues with Bad Use of analytics
 Situation 1: Biased Algorithm (https://www.research.ibm.com/5-in-5/ai-and-bias/)

This is one major headache for data scientists, with possible ramification of lawsuits
Eg 1: Amazon’s hiring algorithm demonstrated gender bias while making hiring decisions
(https://www.reuters.com/video/?videoId=OV91HI8I3&jwsource=cl)
Eg 2: Marriott hotels created an algorithm for managing house-keeping shifts – It resulted
in a union dispute because workers were treated unfairly
(https://www.magzter.com/articles/12425/343398/5cbeb3546fd7f)
▪ Situation 2: Over-reacting to ML predictions
Basically falling into the self-fulfilling prophecy trap
Eg: Suppose an ML algorithm predicts flight risk score for an employee to be very high, a
manager may assume that this employee is a lost cause and behave in a passive
manner, thus triggering the employee’s intention to quit
▪ Situation 3: Abusing privacy data
Nowadays, firms can capture all type of data ranging from social media posts to health-
related metrics. If such data gets disclosed in any public platform ,there can be legal
issues such as breach of Health Insurance Portability and Accountability Act (HIPAA)
How Amazon’s HIRING
ALGORITHM got it wrong
Source: Dastin, J. 2018.

Amazon scraps secret AI
recruiting tool that showed
bias against women. Available
at:
https://www.reuters.com/articl
e/us-amazon-com-jobs-
automation-insight/amazon-
scraps-secret-ai-recruiting-tool-
that-showed-bias-against-
women-idUSKCN1MK08G
Some major goof-ups
 Recent research has uncovered large gender and racial bias in AI systems sold by tech
giants like IBM, Microsoft, and Amazon.
Copyright: Prof. Pratyush Banerjee, IMI Bhubaneswar
Copyright: Prof. Pratyush Banerjee, IMI Bhubaneswar
Source: https://time.com/5520558/artificial-intelligence-racial-gender-bias/
46 The hidden wall of Analytics Maturity
Why can’t most

firms go beyond
level 2?
How to break
this wall?
 Knowledge needed:
 Statistics (correlation,
regression, decision tree,
cluster analysis, factor
analysis)
 Machine learning (logistics
regression, decision tree,
neural networks, random
forests), NLP (Text mining,
sentiment mining)
 Linear programming
Some myths which hold firms back
Myth 1: Software tools are difficult to learn, lots of coding involved

Reality – most applications can be executed through GUI tools
Myth 2: GUI tools are expensive.

Reality – there is GUI interface for both R and Python which are free software
Myth 3: The statistics and Machine learning concepts are very difficult
Reality – can be approached in a simpler way.
So, what do I mean by making analytics
easier to digest?
 What can be ignored at the foundational phase:

a) Coding in R and Python
b) Deep knowledge of statistics
c) Deep knowledge in linear Algebra or Calculus
▪ What can not be ignored:

a) Fundamentals of statistics
b) Fundamentals of machine learning
c) Learning to interpret results in correct manner
d) Understand R / Python in code-agnostic manner (though choice of code-
agnostic software is yours)
Software knowledge required for breaking
the wall
For Descriptive Analytics (Easily Accessible):
MS Excel (Not free but accessible)
Power BI (Available as a part of latest MS Excel suites)
Tableau (Free for students and academicians)
For Predictive Analytics (Open Source / Easily Accessible):

MS Excel (for basic applications, advanced applications require Macros)
R (GUI Packages R Commander for statistics and RATTLE for data mining)
Anaconda Python (GUI Package Orange) – for data mining. Also useful tool
for data visualization
Hardware Requirements – At least 8 GB RAM, Processing speed of 2 GHZ or

more gives faster results and less downtime
Coding knowledge – not required up to a very high level of application
FRAMEWORKS
FOR EXECUTING
ANALYTICS
PROJECTS
The LAMP
Framework
Source: Cascio, W., & Boudreau, J. (2010). Investing in people:

Financial impact of human resource initiatives. Ft Press.
www.sambhavkadam.org/courses-2
https://milvest.sambhavkadam.org
The Mondore, Douthitt and Carson Framework
Source: Scott Mondore, Shane Douthitt, and Marisa Carson. ‘Maximizing the Impact and Effectiveness of HR
Analytics to Drive Business Outcomes,’ People and Strategy 34, no. 2 (2011): 20.
THE LIGHT FRAMEWORK
Source: Banerjee, P. (2021). Predicting Flight Risk of

Employees with Deep Learning. In Press.
54
THANK YOU

Session 1 - Introduction To Data Analytics

Uploaded by

Copyright:

Available Formats

Session 1 - Introduction To Data Analytics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Session 1 - Introduction To Data Analytics

Uploaded by

Copyright:

Available Formats

TRAINING PROGRAM

• Present Designation- Associate Professor–IMI

❑ Business Analytics may be defined as the application of

Most major fortune 500 organizations are

Understanding appropriate data and metrics

Building appropriate platform for data Analytics

In 2004, Filippo Passerini took over as CIO of P&G in 2004. He renamed

Supporting Real-Time Decision-Making through “Decision Cockpits”:

 Netflix collects information on how a user interacted and

 Allrecipes is the world’s largest digital food brand, serving a global

Descriptive analytics: It encompasses the set of

Examples - data queries, reports, descriptive

A Social Media Engagement Dashboard

Statistical Applications- Cross-tabulation, frequency distribution, means

Machine learning Applications:

Performance Group employees by performance and identify Cluster Analysis

Business planning Identify impact of strategic alternatives Correlation and

• Deloitte’s new performance management process requires every team leader

36 The Geeks Arrive In HR-Case of a

Statistical tests – T Tests and AI & ML Based tests – Logistic

 Situation 1: Biased Algorithm (https://www.research.ibm.com/5-in-5/ai-and-bias/)

Source: Dastin, J. 2018.

Copyright: Prof. Pratyush Banerjee, IMI Bhubaneswar

Why can’t most

Myth 1: Software tools are difficult to learn, lots of coding involved

Myth 2: GUI tools are expensive.

 What can be ignored at the foundational phase:

▪ What can not be ignored:

For Predictive Analytics (Open Source / Easily Accessible):

Hardware Requirements – At least 8 GB RAM, Processing speed of 2 GHZ or

Source: Cascio, W., & Boudreau, J. (2010). Investing in people:

Source: Banerjee, P. (2021). Predicting Flight Risk of

You might also like