DATA ANALYTICS

PORTFOLIO
PROJECT
PREPARED BY DEEPAK YADAV

Email Id:[email protected]
PROFESSIONAL BACKGROUND

EDUCATION
Completed B.E - 2018.
Completed Data Analytics Trainee certificate course from Trainity.
Currently learning Python & Tableau

HARD SKILLS ABOUT ME

Structured Query Language (SQL) Thorough and meticulous Data Analyst
Microsoft Excel passionate about helping businesses
Python Programming Language succeed. Looking forward to work in an
Data Visualization (Power BI & Tableau) esteemed company where skills,
Statistics knowledge, talents matter more than the
degree. Through development of portfolio
projects and internship, I learned the
importance of having an iterative,
SOFT SKILLS hypothesis oriented approach to analysis.
Storytelling I am eager to leverage that approach at
Communication Skills your company as a data analyst.
Critical Thinking
Teamwork Working as a HR from last 1.5 years
Research
Presentation and Report

OTHER SKILLS
Predictive Analytics
Risk Analytics
Behavioral Analytics
 TABLE OF CONTENTS
Topics Projects Page

Project description
Data Analytics 6 steps process 4
Process Project Link

Project description
Instagram User Findings & Approach
Analytics Insights 5
Project Link

Operation Project description

Analytics & Findings & Approach
Investigating Insights 6-7
Metric Spikes Project Link

Project description
Hiring Process Findings & Approach
Insights 8
Analytics
Project Link

Project description
IMDB Movie Findings & Approach
Analysis Insights 9
Project Link

Project description
Business Understanding
Bank Loan Case
Findings & Approach 10-11
Study
Insights
Project Link

Project description
XYZ Ads Findings & Approach
Insights 12-13
Airing Reports
Project Link

Project description
ABC Call Findings & Approach
Volume Trend Insights 14-15
Project Link

Learning
Conclusion 16
Portfolio Link
 Data Analytics Process

We use Data Analytics in everyday life.

Description:
For example: Searching in Web Searching Engines like Google.

6 Steps Process:

Plan:
We first decide which things we need to search before opening the Google. Which
information do we need to search? For example, we want to know about different
machine learning algorithms.

Prepare:
Next we need to check which website will give us the correct information in the simplest way.

Process:
Then we need to check how much we want from the data. Like if we want –
difference between the different algorithms / which algorithm is the best choice
according to the different scenarios / etc.

Analyze:
We then analyze the different algorithms. Suppose we are working on some project and
we don’t know how to apply the algorithms or which algorithm will be the best choice in
that project.

Share:
Now we search it in the Google. And Google gives us the best results at the top-most by
using the process of data analytics / data science.

Act:
Then we finally click it and get the necessary information which we need. We, then use
that information in our project.

Project Link:
https://drive.google.com/file/d/1rB0VmFRLSrMd
RUciMiPcUqBf02QRsBnW/view?usp=sharing
 Instagram User Analytics
This project is about how the users engage and interact with Instagram.
Description:
We will analyse these users in an attempt to derive business insights for
marketing, product & development teams. These insights are then used
by teams across the business to launch a new marketing campaign,
decide on features to build for an app, track the success of the app by
measuring user engagement and improve the experience altogether while
helping the business grow.

Rewarding Most Loyal Users: The 5 oldest users of the Instagram

Findings:
Remind Inactive Users to Start Posting: The users who have
never posted a single photo on Instagram
Declaring Contest Winner: The user who gets the most likes on a
single photo
Hashtag Researching: The top 5 most commonly used hashtags
Launch AD Campaign: Day of the week do most users register on
User Engagement: Average user posts on Instagram
Bots & Fake Accounts: Users (bots) who have liked every single photo
on the site

We are working with the product team of Instagram and the product
Approach:
manager has asked us to provide insights on the questions asked by the
management team. We use SQL to derive different insights from the
dataset provided by the management team. First, we run the necessary
commands for creating the database to work on. Then, we performed
analysis to generate valuable insights for the company.

Insights: There are total of 100 users using Instagram clone.

Around 26% of the users are inactive in Instagram. We can remind the
inactive users by sending them promotional emails to post their 1st
photo.
The most liked photo in Instagram is posted by Zack_Kemmer93,
which is liked by 48% of the users. The team can start the contest for
the most liked photos. This will make the users to post more such
good posts.
The most used hashtag is “smile”. Around 59% of the users use the
“smile” hashtags. If a partner brand use the “smile” hashtag, it will be
able to reach the most users in the platform.
The best days to launch ads are Sunday and Thursday. As the most
users register on Instagram on Sunday and Thursday.

Project Link: 13% of Instagram IDs are fake and dummy accounts.
https://docs.google.com/document/d/1XEbjny2iJtzAn3jI4lvf8CguvRJgcq-N/edit?usp=sharing&
ouid=109059427539152145245&rtpof=true&sd=true
Operation Analytics & Investigating Metric Spikes
Operation Analytics is the analysis done for the complete end to end
Description:
operations of a company.With the help of this, the company then finds
the areas on which it must improve upon.
Being one of the most important parts of a company, this kind of analysis
is further used to understanding between cross-functional teams, and
more effective workflows.

Investigating metric spike is also an important part of operation analytics

as being a Data Analyst we must be able to understand or make other
teams understand questions like- Why is there a dip in daily engagement?
Why have sales taken a dip? Etc. Questions like these must be answered
daily and for that it’s very important to investigate metric spike.

Number of jobs reviewed

Findings:
Throughput
Percentage share of each language
Duplicate rows
User Engagement
User Growth
Weekly Retention
Weekly Engagement
Email Engagement

Approach: I am working for a company like Microsoft designated as Data Analyst

Lead and is provided with different data sets, tables from which I must
derive certain insights out of it and answer the questions asked by
different departments.
Firstly, I spent some time on understanding the data/table given. I
cleared the questions which was in my mind and what are the things to
consider while reviewing the data. I use SQL to derive different insights
from the dataset provided by the management team. I first created a
database “operation_analytics” and then the tables using the structure
and links provided by the team. Then, we performed analysis to generate
valuable insights for the company.

Insights: 1. Case Study 1 (Job Data):

The number of distinct jobs reviewed per hour per day for November
2020 is 83%.
We used the 7-day rolling average of throughput as it gives the average
for all the days right from day 1 to day 7 whereas, daily metric gives the
average for only that particular day itself.
The percentage share of Persian language is the most (37.5%).
There are two duplicate rows if we partition the data by job_id. But if we
look the overall columns, all the rows are unique.
 2. Case Study 2 (Investigating metric spike):

The weekly user engagement increased from week 18th to week 31st
and then started declining from then onward. This means that some of
the users do not find much quality in the product/service in the last of
the weeks.
There are in total 9381 active users from 1st week of 2013 to the 35th
week of 2014.
The overall count of weekly engagement per device used is the most
for MacBook users and iPhone users.
The email opening rate is around 34% and email clicking rate is around
15%. The users are engaging with the email service which is good for
the company to expand.

Project Link: https://docs.google.com/document/d/1h4fbAuQxJttFcEG-Io-KopgmaI57O

Y5J/edit?usp=sharing&ouid=109059427539152145245&rtpof=true&sd=tr
ue
 Hiring Process Analytics

Hiring process is the fundamental and the most important function of

Description:
a company. Here, the MNCs get to know about the major underlying
trends about the hiring process. Trends such as- number of rejections,
number of interviews, types of jobs, vacancies etc. are important for a
company to analyse before hiring freshers or any other individual.
Being a Data Analyst, our job is to go through these trends and draw
insights out of it for hiring department to work upon.

Findings: Hiring: How many males and females are Hired ?

Average Salary: What is the average salary offered in this company ?
Class Intervals: Draw the class intervals for salary in the company ?
Charts and Plots: Draw Pie Chart / Bar Graph ( or any other graph ) to
show proportion of people working different department ?
Charts: Represent different post tiers using chart/graph?

I am working for a MNC such as Google as a lead Data Analyst and the
Approach:
company has provided with the data records of their previous hirings
and have asked me to answer certain questions making sense out of
that data.
We will use EDA to generate different insights and to answer the
questions asked by the company.
The dataset given by the company contains the details about people
who registered for a particular post in a department of this company. I
used MS Excel to analyze the data with different tables and columns.

The rejection rate of male applicant is 6% higher than the female

Insights:
applicant.
The average salary paid in this company is 50K.
Most of the employers are in the Operation Department and then in
the Human Resource Department.
The applicant is most likely to get hired if he/she is applying for the
HR Department as the rejection rate here is the least.
There are only 3 candidates in the company who are paid more than
100K.

Project Link: https://docs.google.com/document/d/1B4yxuWUY3n43BGXnkNO6Fo

uZvnOlCjBk/edit?usp=sharing&ouid=109059427539152145245&rtpof
=true&sd=true
 IMDB Movie Analysis

The dataset provided by the company contains various columns of

Description:
different IMDB Movies. We are required to Frame the problem. For
this task, we will need to define a problem we want to shed some
light on.
We can do this by asking the following 'What?' :
What do we see happening?
What is our hypothesis for the cause of the problem? (this will be
broadly based on intuition initially)
What is the impact of the problem on stakeholders?
What is the impact of the problem not being solved?

The things that we find out through the project are:

Findings:
movies with the highest profit
top movies as per imdb rating
top directors
most popular genres
top foreign language films

Firstly, I cleaned the data. Then, we used Five 'Whys' approach to

Approach:
determine its root cause by repeatedly asking the question “Why”.
While asking Why is easy, what we're interested in is the answer.
Each time we answer why the next time gets more difficult as we
must think deeper behind the reasons for this. As we ask why, we
may find that we have multiple answers for the same question.
Insights: There are as many as 5 outliers in the profit columns.
The movie with the highest profit is ‘Avatar’ followed by ‘Jurassic
World’ and ‘Titanic’ and so on.
The Shawshank Redemption is the top-most movie with the highest
IMDB rating.
The Good, the Bad and the Ugly (Italian) is the top-most foreign
language movie.
Charles Chaplin is the top-most director followed by Tony Kaye.
The most popular genres is Drama followed by Comedy.
'Leonardo DiCaprio' is the critic-favorite as well as the audience-
favorite actor.
The most users voted in the decade 2000s and the least in the
decade 1940s.

https://docs.google.com/presentation/d/1rVQl-934cwQqaRU6lIpw3Epx
Project Link:
LdcTN-8D/edit?usp=sharing&ouid=109059427539152145245&rtpof=tr
ue&sd=true
 Bank Loan Case Study

This case study aims to give us an idea of applying EDA in a real

Description:
business scenario. In this case study, we will develop a basic
understanding of risk analytics in banking and financial services and
understand how data is used to minimize the risk of losing money
while lending to customers.

The loan providing companies find it hard to give loans to the

Business
people due to their insufficient or non-existent credit history.
Understanding: Because of that, some consumers use it as their advantage by
becoming a defaulter.
Suppose we work for a consumer finance company which
specialises in lending various types of loans to urban customers. We
have to use EDA to analyse the patterns present in the data. This will
ensure that the applicants capable of repaying the loan are not
rejected.

Our aim is to identify the patterns which indicate if a client has

Findings:
difficulty paying their installments which may be used for taking
actions such as denying the loan, reducing the amount of loan,
lending (to risky applicants) at a higher interest rate, etc.
The driving factors (or driver variables) behind loan default, i.e. the
variables which are strong indicators of default.
Presenting the overall approach of the data analysis, cleaning the
dataset, finding outliers, data imbalance, univariate, segmented
univariate, bivariate analysis, etc.
The top 10 correlation for the Client with payment difficulties and all
other cases (Target variable).

Imported the NumPy, pandas, matplotlib and seaborn python

Approach:
libraries.
Imported the datasets (Application_Data & Previous_Application)
Identification: We have identified how we will approach the data,
finding missing dataset and working on it accordingly to gain the
required results.
Outliers: Identified outliers and showed how they play if any role
in our dataset.
Imbalance: Understanding the ratio of imbalance in our data.
Correlation Analysis: Finding the correlation between the variables
with respect to the target variables and find the top three
correlation.
Visualisation: Visualized the data with the help of charts and
graphs.
 Insights: 1. Non-Default:
Academic degree has less defaults.
Student and Businessmen have no defaults.
Clients with Trade Type 4 and 5 and Industry type 8 have defaulted
less than 3%.
People above age of 50 have low probability of defaulting.
Applicant with Income more than 700,000 are less likely to default.
People with zero to two children tend to repay the loans.

2. Default:
Men are at relatively higher default rate.
Clients who are either at Maternity leave OR Unemployed default a
lot.
Not approving the loan of young people who are in age group of 20-40
as they have higher probability of defaulting.
When the credit amount goes beyond 3M, there is an increase in
defaulters.
People who have less than 5 years of employment have high default
rate.

https://drive.google.com/file/d/1j-GELtBkl6Y50GuImk0uTtAzMTbVTCbc
Project Link:
/view?usp=sharing
 XYZ Ads Airing Report Analysis

Advertising is a way of marketing our business in order to increase

Description:
sales or make our audience aware of our products or services. Until a
customer deals with us directly and actually buys our products or
services, our advertising may help to form their first impressions of
our business. Target audience for businesses could be local, regional,
national or international or a mixture. So they use different ways for
advertisement. Some of the types of advertisement are:
Internet/online directories, Trade and technical press, Radio, Cinema,
Outdoor advertising, National papers, magazines and TV. Advertising
business is very competitive as a lot of players bid a lot of money in a
single segment of business to target the same audience. Here comes
the analytical skills of the company to target those audiences from
those types of media platforms where they convert them to their
customers at a low cost.
What is Pod Position? Does the Pod position number affect the
Findings:
amount spent on Ads for a specific period of time by a company?
What is the share of various brands in TV airings and how has it
changed from Q1 to Q4 in 2021?
Conduct a competitive analysis for the brands and define
advertisement strategy of different brands and how it differs
across the brands.
Mahindra and Mahindra wants to run a digital ad campaign to
complement its existing TV ads in Q1 of 2022. Based on the data
from 2021, suggest a media plan to the CMO of Mahindra and
Mahindra. Which audience should they target?

Scatter chart w.r.t different brands is used to know if the pod

Approach:
position affects the amount spent on Ads for a specific period
of time by the company.
Bar chart and column chart are used for answering the share of
various brands in TV Airings.
We used pivot table to conduct the competitive analysis for
the brands.
Clustered Column chart is used to suggest a media plan to the
CMO of Mahindra and Mahindra
 The brand’s money spent for the advertisement is the least for the
Insights:
last quarter pod position and the highest for the first quarter pod
position.
The money spent by Mahindra and Mahindra is the most for the pod
position ads.
The money spent by Honda Cars is the least for the pod position ads.
The money spent by the Maruti Suzuki is the most consistent for all
the Quarters of the year.
People watch the most in the prime time and on weekend.
The Ads are shown the least in the prime access and evening news
parts of the day.

https://drive.google.com/file/d/17Azv3yhCuM2DhOQCzsm_goaIzmM2l0Y
Project Link:
F/view?usp=sharing
 ABC Call Volume Trend Analysis
The attached dataset is of Inbound calls of an ABC company from
Description:
the insurance category consists of a Customer Experience (CX)
Inbound calling team for 23 days. Data includes Agent_Name,
Agent_ID, Queue_Time [duration for which customer have to wait
before they get connected to an agent], Time [time at which call was
made by customer in a day], Time_Bucket [for easiness we have also
provided you with the time bucket], Duration [duration for which a
customer and executives are on call, Call_Seconds [for simplicity we
have also converted those time into seconds], call status (Abandon,
answered, transferred).

The average call time duration for all incoming calls received by
Findings:
agents (in each Time_Bucket).
The total volume/ number of calls coming in via charts/ graphs
[Number of calls v/s Time].
Propose a manpower plan required during each time bucket
[between 9am to 9pm] to reduce the abandon rate to 10%.
Propose a manpower plan required during each time bucket in a
day[9 pm to 9 am]. Maximum Abandon rate assumption would be
same 10%.
Approach: We used pivot table and pivot charts to get the valuable
insights of the data.
We assumed an agent work for 6 days a week;
On an average total unplanned leaves per agent is 4 days a
month; An agent total working hrs is 9 Hrs out of which 1.5 Hrs
goes into lunch and snacks in the office.
On average an agent occupied for 60% of his total actual
working Hrs (i.e. 60% of 7.5 Hrs) on call with customers/ users.
We also assumed total days in a month is 28 days for easy
calculation.
 The customers call the least in the evening. So, the company can
Insights:
reduce the number of agents at that time for answering the calls.
The company can hire 17 customer support agents for the night shift
work.
The company can shift some of the day workers for the night shift.
The employees who are working 9 am to 9 pm. The manager can
change some of the workers shift from 5 am to 2 pm and some
workers from 2 pm to 11 pm to get the most calls answered.
The company can make the employers divide into 3 parts too, so that
the agents are always available 24/7.

Project Link:
https://drive.google.com/file/d/13BVWMxiomQkblAg0ZfgJKiiy2rUBwh88/vi
ew?usp=sharing
 Conclusion
The things I learned from the projects are:
Learnings:
Data Analysis 6 Steps Processes
How to use Advance SQL Concepts in the real world
business case
How to use Advance Excel Concepts in the real business
case scenario
How to analyze the huge datasets in Python, Excel, etc.
How to visualize the data to gain the valuable insights
Concepts of Operation Analysis & Investigating Metric
Spikes
HR Analytics
Predictive Analytics
Risk Analytics
Behavioral Analytics
Business scenario of Ads Airing Report
Customer Experience Team and Inbound Customer Support
Google to get the concepts and answers whenever get stuck

Contact: [email protected]

Thank You