1.

INTRODUCTION

SYNOPSIS:

Purchasing products is a part of our daily life. Even though the online shopping has
created a good impact on people, the retail shops do not fail to prove it’s importance to the
consumers. Due to the evolvement of online shopping, it is necessary for the retail shop
holders to improve their sales mechanism by providing some attractive offers to the
consumers. The retailers should make sure that the end users (i.e) customers do not face
much difficulties in purchasing products. Retail industry contributes a considerable part to
our economy as well. Their contribution will be in the form of producing, distributing, and
selling the goods you use on a daily basis like food, clothes, fuel, and so on . The main
objective of this project is to identify the intercept value of total sales and total quantity,
Products that are frequently purchased together and to identify if the factors such as
consumer’s income, region, owning car, preference of online portal for shopping etc have any
influence on their purchasing behavior. For analyzing these, machine learning technique like
linear module, frequency pattern growth and classification method like k- means clustering is
used.

OBJECTIVES:

 To predict total sales.

 To predict total quantity.
 To identify the products which are frequently purchased together?
 To segment the customer ID based upon the factor of preference of weekly
purchases.
 To identify the relationship between factors such as marital status, owning of car,
number of children, region etc and their purchasing pattern.

1
1.1.COMPANY PROFILE:

Kartech provide guidance to students and scholars in their M.Phil and Ph.D research
and project work. We help individuals, professionals, students, researchers to compete with
the cutting edge technologies. In simple Kartech is a research, development and technological
company to empower talents across computer, management, biological and statistical studies
to achieve great heights.

Project development:

At Kartech we plan, design, and deliver solutions to complex problems of client’s

project and help minimize the delivery time and production cost. We know that turning
organizational goals into results is quiet challenging; Kartech help clients achieve their
business challenges through successful change management through business analysis
knowledge and utilizing technological support.

Business Situations:
 Deployment, Operation and Enhancement
 Project Review
 Project Methodology implementation
 Demand Management
 Resource Management
 Vendor Management
 Project planning management
 Business architecture
 Application management
 Data Management & optimization
 Clients Relationship
 Survey & Research
 Sales Lead Generation
 Branding

2
Business Analytics:
Using multiple combinations of financial analysis, statistics,
research, funneling and target segment we analyze and develop best solutions to companies.
Our unique approach and market segmentation has derived positive factors that play a major
role in sales, Marketing and branding.
Using industry expertise and relevant experience gained through
years, we have a unique way of understanding the metrics of analytical data into insights and
helping every product, brand to obtain a clear sales growth. Using the insights clients can
understand the market trends, buying patterns and give customers a greater experience over
brands & products and services.

IEEE Project Works:

We assist in implementing your IEEE Current Papers, or
alternatively help you choose your paper according to your preferred domain. We deal with
Cloud Computing, Data Mining, Networking, Security, Android, Image Processing, Project
implementation etc.
Our Features:
 E-Commerce solutions
 Automated solutions to enable business growth without increasing head count
 User friendly, dynamic, web-enabled productivity tools for analysis
 Integrated data across platforms, data sources and locations
 Improved reporting and query capabilities
 High degree of control over data security
 Enhanced target marketing efforts.
 Increase in customer retention and cross sales.
 Increase in customer satisfaction.
 Increase in new business Markets.
 Increase in response rates from marketing campaigns.
 Increase in return on investment.

3
1.2.ABOUT THE PROJECT:

Retail industry simply means sale of finished goods and services by shopkeepers (or)
online retailers to end user or customers. The chain always start from the producer and ends
with retailers. Manufacturers create the product. The middle of the supply chain is wholesale
sales. They distribute the goods and services to retailers. The retailers sell them to the
consumer. The most critical time in retail sales is the holiday shopping season. It accounts for
almost 20 percent of annual sales for many retailers.

The Indian retail industry has shown immense growth during past five years. India is
the most preferred retail destination worldwide, because it satisfies the most important factor
which is availability of consumers in mass. Retail industry reached to Rs 66.39 lakh crore in
2018 and expected to reach Rs 76.87 lakhs crores by the end of 2020. The major contribution
is nowadays from e-retailing. Though online shopping is preferred mostly, retail shops never
fail to prove it’s importance customers’ daily life need. They use various machine learning
techniques to analyze the behavior of the customers from where they see when entering the
shop till the products they choose while billing and arrange the products accordingly.

Machine learning is an application of artificial intelligence (AI) that provides systems

the ability to automatically learn and improve from experience. Machine learning focuses on
the development of computer programs that can access data and use it learn for themselves.
This data consist of Bill no. of customers, their gender, unit cost of products, total cost of
products product name, consumer preference of online shopping, customers’ marital status,
customers’ income level etc,. Their purchasing behavior is analyzed using machine learning
techniques such as linear model and frequency pattern growth.

THE PROJECT CONSISTS OF THE FOLLOWING MODULES:

 Product Name
 Product ID
 Bill Number
 Customer’s Gender
 Unit Cost
 Total Quantity
 Total Sales

4
CORRELATION:
 Relationship between Weekly purchase and gender
 Relationship between Weekly purchase and customer’s income
 Relationship between Weekly purchase and region
 Relationship between Weekly purchase and promotional mails
 Relationship between Weekly purchase and Preference of online portal for shopping.

5
 2.REVIEW OF LITERATURE

With a population of 108 billion and growing at 1.6% p. a, the retail sector in India
has a huge growth potential to become one of the leading retail markets of the world. With a
young demographic profile, increasing consumer aspirations and advent of dual income
families, the retail landscape is undergoing a level of transformation. The sector which was
initially dominated by highly decentralized unorganized retailers, with the demographic and
psychographic changes, is becoming centralized and organized as well. Food dominates the
shopping basket in India, accounting for almost 76% of the spending; while it is only 9.7% in
the U.S and 15% in Japan and UK. Out of the total 76%, only 1% of this market is in the
organized sector and the remaining is in the unorganized sector. This research study aims at
finding out the customer preferences towards the organized and unorganized with special
reference to fresh groceries. An attempt has been made to identify factors that influence
customer preferences towards the organized and the unorganized retailers[13] .

Retailing consists of all activities involved in selling goods and services to consumers
for their personal, family, or household use. It covers sales of goods ranging from
automobiles to apparel and food products, and services ranging from hair cutting to air travel
and computer education retailing is one of the largest sectors in the global economy. In India
for a long time the corner grocery store was the only choice available to the consumers with
the increasing demand of the customers spurred by changing trends, aspiring needs for
variety, the traditional retail gave rise to modern retail format. The traditional food and
grocery segment has seen the emergence of supermarkets/grocery chains, convenience stores
and hypermarkets [7].

Private labels entered in Indian retail almost two and half decade before and still have
not picked up as compared to their dominance in western part of the world. With recent
growth of organized retail in India, we can see these private labels as emerging force in
Indian retail. Many of Indian retailers like Future Group, Tata’s Croma, Aditya Birla’s More,
Shoppers Stop, Spencer’s etc are relying on private label strategy in a big way. Besides, rapid
technological and socio-economic changes over the last two decades have affected the buying
behavior of consumers forcing retailers to innovate and build new brands (PLBs/Store
Brands) across different categories to attract more buyers in their stores. This research study

6
the factors affecting the sales growth of private labels in India and also determining and
comparing customer’s demographic profile for Private Label brand and national brand [9].

Since the onset of e-commerce, e-businesses have always been interested in knowing
what propels their customers. An extensive literature review has been cast around the said
reason to identify factors affecting online buying behavior. Twenty-six factors have been
identified initially, out of which price; convenience; security; information; enjoyment; access
and tangibility/ sensation have been found to be most cited. These factors have been further
explored to examine their implications in managerial decision making [11].

The global markets have witnessed major shifts in the consumers’ behavior that have
been much influenced by the change of technology, innovation, research and development.
The consumer’s needs thus dynamically change in order to respond to the change in the
social and business environment. Customers belonging to different age groups flock towards
these outlets to satisfy their taste buds and hedonistic needs. Out of all the age groups, the
youngsters are considered to be the primary target by the fast food service industry. Hence,
these fast food services try to lure its customers through sensory appeals [15].

The retail industry has seen a tremendous growth in the size and market dominance of
larger players, with greater store size, increased retail concentration, and utilization of a range
of formats. With the inception of retailing recent years, various types of sophisticated and
complex formats have originated in the retailing scenario of India. Each type of retailer is
trying to survive and prosper by satisfying a group of customers` needs and expectations
more effectively than its competitors. The retailers have succeeded in their endeavor to a
great extent. The growth of an organized retailing is predominant across the globe. Organized
retail is on the threshold of a boom in India. Branding is an important element in the retailing
industry to influence customer perceptions and to drive store choice and loyalty [8].

The retailing industry is one of the booming industries in India. It is been considered
as equal competent with IT industry. Shri Siddeshwar Super bazaar is one of the retail stores
in Bijapur city. In view of above the survey on to understand the customer behavior towards
Shri Siddeshwar super bazaar. Objectives of the study are to understand the shopping
behavior of customers, to study the factors influencing purchase in super bazaar, to study the
customer satisfaction level towards super bazaar, to compare super bazaar with other retail
stores, to understand the impulse buying nature [14].

7
 Store choice is a decision that a shopper is fairly involved in. It is important for a store
to understand this behavior for developing marketing strategies to attract and keep its
clientele. It is found that shoppers choose the store based on many aspects that could be
classified as primary and image based. It is also found that the importance of each of these
aspects changes with the kind of store the shopper wants to visit. In the Indian context where
the shopper does not have much variety in store format, the type of store is recognized by the
kind of product the store deals in. The paper is an attempt to understand this behavior of the
shopper. The shoppers are explored for the primary reasons for choosing a store. Then, using
a factor analysis, the several image dimensions are classified. Further, using multinomial
logit regression, the store choice pattern is studied across different types of store.
Implications for the managers in the retail business are drawn and future research directions
have been highlighted [10].

A shopping situation is a particular act of buying behavior occurring at a specific

point in space and time. The purpose of this paper is to examine the impact of situational
factors such as task definition, perceived risk, physical surroundings, social surroundings and
temporal aspects on retail format choice decisions relating to kirana stores, convenience
stores supermarkets and hypermarkets in the fast growing food and grocery retailing in India.
The study is purely based on primary data collected from 1040 retail customers through mall
intercept method using structured and non-disguised questionnaire from sixty different food
and grocery stores from twin cities of Secunderabad and Hyderabad in Andhra Pradesh in
India. The multiple discriminant analysis reveals that physical surroundings (ambience, store
design and visual merchandising), task definition (regular purchase, purchase in large
quantities and getting ideas/knowing new products), perceived risk (time, financial, and
physical), temporal aspects (time spent and convenient timing hours), and social interactions
and experiences have significant affect on supermarket and hypermarket store format choice
decisions [4].

Retailing in Malaysia has grown significantly over the years witnessing foreign
retailers with expanded retail brands entered the country, offering a wider choice of selection
to consumers in terms of brands and styles. The retailing landscape continues to gain
sophistication with new retailing formats, products and services being introduced. Overall,
grocery retailers saw an increase in current value terms in 2010. As grocery retailers
especially hypermarkets, started to carry non-grocery products, such as home appliances, they
began to compete with purely non-grocery retailers. With consumers demanding wider

8
choices from retailing, and taking on more international tastes, competition within retailing in
Malaysia intensified [3].

In Indian traditional grocery retailing from the perspective of shoppers having

different family sizes. Grocery is one of the major sectors of retail globally and it would be of
academic as well as business interest to understand demographic peculiarity as seen in the
context of Traditional Grocery retailing. It is hypothesised that shoppers of different family
size would have different perception of various retail outlets as their requirements of grocery
goods are different. Also as Indian grocery retail is dominated by the traditional grocery
stores known as ―Kirana‖; knowledge of the difference in perception of shoppers‘ with
different family size would lead to a better understanding by these stores so as to devise their
strategies to enhance the shoppers‘ patronage [5].

The human race has been performing with high level of regularity and involvement.
Over the years, organized and unorganized retail shoppers’ orientation towards routine
activity has been changing with the inception of organized retail in India. The innovations
brought by retailers and marketers in the practice of retailing have been providing new
paradigms for shopping and consumer loyalty. This has also led to a body of knowledge that
aims to understand orientation of retail shoppers towards consumer loyalty [1].

Online shopping is a growing phenomenon worldwide and with increase in internet

penetration India too is witnessing a surge in online shopping. Consumer buying behavior is
a widely researched area but online buying behavior is a relatively new field of study.
Marketers are trying to understand the online buying behavior of Indian consumers so that
they can look out for new ways of identifying distinct profitable customer segments and
targeting them better. Marketers need to know who are the online customers , what are they
buying online and why, what prevents them from buying online , what other factors affect
their online buying behavior, does social media have any influence on their buying behavior
[6].

On customer evaluations when taken as a comparative between two retail segments.

Two of the largest retail segments viz. grocery & apparels have been taken for the purpose of
comparison of importance according to store attributes. This study is significant given that
the retail sector in India is presently on a growth trajectory & the right retail format is still a

9
very critical area of managerial decision making. The retail format- which represents the right
mix of various store dimensions- will eventually depend upon the interplay between various
store attributes. This study uses a research instrument developed by the author in an earlier
study for the comparative analysis. A very significant pointer from this research is that while
there is some commonality of attributes between retail sectors, the precise importance and
mix are, arguably, determined more by the motivation of the customer behind each specific
shopping excursion. The grocery and apparel store attributes dimensions that have emerged
from this study prove that the factors are different in terms of their composition and
importance.

The retail industry in India is slowly taking a new shape and structure. The dominant
unorganized retailing started losing its stand in the market and the organized sector is taking
its importance in the industry. The organized retailers have changed the perceptions of the
consumers and chased out the consumer from the unorganized retailers. What the customers
look for from today’s retailers is ambience, price, variety, assortments and merchandise
which are addressed meticulously by the organized retailers. With a change in the purchasing
power, tastes and preferences of the Indian retail consumer, read availability of product with
the retailer is playing a major role in the sustainability of the retailers. In this context supply
chain management (SCM) plays a vital role. The mechanism which helps in decreasing the
waste, decreasing the inventory, decreasing the transportation cost, finally giving, the
consumer an affordable price and improving the performance of the retailer is what is called
as Supply Chain Management.

10
 3.SYSTEM ANALYSIS

3.1.PROPOSED SYSTEM ANALYSIS:

The primary aim of this project is to identify the accurate intercept relationship
between total cost, total sales and total quantity; to identify the products that are frequently
purchased together in order to make some attractive offers for the customers, to identify the
class of people who prefer weekly purchases and to identify if the gender, region, number of
children, owning of car etc have any influence in purchasing behavior of the customers.
Linear regression is used to find the accurate intercept value of Total cost and Total quantity.
It is a common Statistical Data Analysis technique. It is used to determine the extent to which
there is a linear relationship between a dependent variable and one or more independent
variables. Frequency pattern growth is used to identify the products which are frequently
purchased together. And Chi Square test is used to identify the relationship between gender,
income and purchasing pattern.

11
 4.FEASIBILITY STUDY

FEASIBILITY STUDY:

It is important for the every retail shop keepers to improve their sales mechanism in
order to gain more number of customers. This is due to the emergence of online shopping
portals. They also have to know the accurate quantity and sales price in order prove justice to
their loyal customers.

In order to make the customers stay loyal to a particular shop they will have to
analyze their purchasing pattern accurately. They may like different products of same brand.
It will be more attractive if they put combination of all those products. With this, the
shopkeepers may get a chance to sell those products which is not much popular in that
particular brand. So this is beneficial for both customers as well as the shopkeepers.

Apart from these, identifying the particular income class who would prefer weekly
shopping is more important. Only then the shopkeepers may choose products for discounts
and much more attractive weekend offers. The prediction of this is very important to be
accurate because if the products they have chosen goes wrong, the customers will think that
the product the shopkeepers have chosen is too much for their income level.

12
 5.SYSTEM SPECIFICATION

5.1.HARDWARE REQUIREMENTS:

 Edition: Windows 10 enterprise

 Version: 1703
 Installed ram: 4 GB
 System Type: 64- bit operating system, x64- based processor
 Processor: Intel Core i3

5.2.SOFTWARE REQUIREMENTS:

 Operating System - Windows 10

 Front-End Tool - R-Studio

5.2.1.ABOUT THE SOFTWARE:

R PROGRAMMING LANGUGAE:

R is a programming language and free software environment for statistical computing

and graphics that is supported by the R Foundation for Statistical Computing. The R language
is widely used among statisticians and data miners for developing statistical software and
data analysis. Polls, surveys of data miners, and studies of scholarly literature databases show
that R's popularity has increased substantially in recent years. Computational statistics, or
statistical computing, is the interface between statistics and computer science. It is the area of
computational science (or scientific computing) specific to the mathematical science of
statistics. This area is also developing rapidly, leading to calls that a broader concept of
computing should be taught as part of general statistical education. R ranks 8th in the TIOBE
index.

13
 R provides a wide variety of statistical (linear and nonlinear modelling, classical
statistical tests, time-series analysis, classification, clustering, …) and graphical techniques,
and is highly extensible. The S language is often the vehicle of choice for research in
statistical methodology, and R provides an Open Source route to participation in that activity.

One of R’s strengths is the ease with which well-designed publication-quality plots
can be produced, including mathematical symbols and formulae where needed. Great care has
been taken over the defaults for the minor design choices in graphics, but the user retains full
control.

R is available as Free Software under the terms of the Free Software

Foundation’s GNU General Public License in source code form. It compiles and runs on a
wide variety of UNIX platforms and similar systems (including FreeBSD and Linux),
Windows and MacOS.

R is a GNU package. The source code for the R software environment is written
primarily in C, Fortran, and R. R is freely available under the GNU General Public License,
and pre-compiled binary versions are provided for various operating systems. While R has a
command line interface, there are several graphical front-ends available.

The R programming language includes functions that support linear modeling, non-
linear modeling, classical statistics, classifications, clustering and more. It has remained
popular in academic settings due to its robust features and the fact that it is free to download
in source code form under the terms of the Free Software Foundation's GNUgeneral public
license. It compiles and runs on UNIX platforms and other systems including
Linux, Windows and macOS.

THE R ENVIRONMENT

R is an integrated suite of software facilities for data manipulation, calculation and

graphical display. It includes an effective data handling and storage facility, a suite of
operators for calculations on arrays, in particular matrices, a large, coherent, integrated
collection of intermediate tools for data analysis, graphical facilities for data analysis and
display either on-screen or on hardcopy, and a well-developed, simple and effective

14
programming language which includes conditionals, loops, user-defined recursive functions
and input and output facilities.

The term “environment” is intended to characterize it as a fully planned and coherent

system, rather than an incremental accretion of very specific and inflexible tools, as is
frequently the case with other data analysis software.

The R language programming environment is built around a standard command-line

interface. Users leverage this to read data and load it to the workspace, specify commands
and receive results. Commands can be anything from simple mathematical operators,
including +, -, * and /, to more complicated functions that perform linear regressions and
other advanced calculations.

Users can also write their own functions. The environment allows users to combine
individual operations, such as joining separate data files into a single document, pulling out a
single variable and running a regression on the resulting data set, into a single function that
can be used over and over.

Looping functions are also popular in the R programming environment. These

functions allow users to repeatedly perform some action, such as pulling out samples from a
larger data set, as many times as the user wants to specify.

R, like S, is designed around a true computer language, and it allows users to add
additional functionality by defining new functions. Much of the system is itself written in the
R dialect of S, which makes it easy for users to follow the algorithmic choices made. For
computationally-intensive tasks, C, C++ and Fortran code can be linked and called at run
time. Advanced users can write C code to manipulate R objects directly.

Many users think of R as a statistics system. We prefer to think of it of an

environment within which statistical techniques are implemented. R can be extended (easily)
via packages. There are about eight packages supplied with the R distribution and many more
are available through the CRAN family of Internet sites covering a very wide range of
modern statistics.

R has its own LaTeX-like documentation format, which is used to supply

comprehensive documentation, both on-line in a number of formats and in hardcopy.

R AND BIG DATA

15
 These limitations have mitigated the applicability of the R language in big data
applications. Instead of putting R to work in production, many enterprise users leverage R as
an exploratory and investigative tool. Data scientists will use R to run complicated analyses
on sample data and then, after identifying a meaningful correlation or cluster in the data, put
the finding into product through enterprise-scale tools.

Several software vendors have added support for the R programming language to their
offerings, allowing R to gain a stronger footing in the modern big data realm. Vendors
including IBM, Microsoft, Oracle, SAS Institute, TIBCO and Tableau, among others, include
some level of integration between their analytics software and the R language. There are also
R packages for popular open source big data platforms, including Hadoop and Spark.

R excels at creating and running Monte Carlo simulations, and the NPS simulation
described above took a mere nine lines of code. I would love to tell you that I'm a hero
because I managed to do it in nine lines, but that really isn't the case. The R programming
language is simply exceptionally good at generating huge sets of numbers and then
manipulating them. It's also good for prototyping big data manipulations.

STATSICAL FEATURE

R and its libraries implement a wide variety of statistical and graphical techniques,
including linear and nonlinear modeling, classical statistical tests, time-series analysis,
classification, clustering, and others. R is easily extensible through functions and extensions,
and the R community is noted for its active contributions in terms of packages. Many of R's
standard functions are written in R itself, which makes it easy for users to follow the
algorithmic choices made.R as a general purpose programming language. I liked Python
before because Python has rich built-in types like sets, dicts, lists, tuples. These structures
help write short scripts to process data. ... But the idea of using R as a general purpose
language is still vague For computationally intensive tasks, C, C++, and Fortran code can be
linked and called at run time. Advanced users can write C, C++, Java, .NET or Python code
to manipulate R objects directly. R is highly extensible through the use of user-submitted
packages for specific functions or specific areas of study. Due to its S heritage, R has stronger
object-oriented programming facilities than most statistical computing languages. Extending
R is also eased by its lexical scoping rules.

16
 Another strength of R is static graphics, which can produce publication-quality
graphs, including mathematical symbols. Dynamic and interactive graphics are available
through additional packages.

R has Rd, its own LaTeX-like documentation format, which is used to supply
comprehensive documentation, both on-line in a number of formats and in hard copy.

6.PROJECT DESIGN

6.1.PROJECT DESCRIPTION:

Retail industry simply means sale of finished goods and services by shopkeepers (or)
online retailers to end user or customers. The chain always start from the producer and ends
with retailers. Manufacturers create the product. The middle of the supply chain is wholesale
sales. They distribute the goods and services to retailers. The retailers sell them to the
consumer. The most critical time in retail sales is the holiday shopping season. It accounts for
almost 20 percent of annual sales for many retailers.

This data comprises of product name, Product ID, Unit cost, Total cost, income of
customers, gender, and marital status, preference of online shopping and preference of week
end shopping. The main objective is to identify the behavior of the customers and to predict
the accurate intercept value of sales and quantity.

The Indian retail industry has shown immense growth during past five years. India is
the most preferred retail destination worldwide, because it satisfies the most important factor
which is availability of consumers in mass.

THE PROJECT CONSISTS OF THE FOLLOWING MODULES:

 Product Name

17
  Product ID
 Bill Number
 Customer’s Gender
 Unit Cost
 Total Quantity
 Total Sales

CORRELATION:
 Relationship between Purchasing pattern and gender
 Relationship between Purchasing pattern and customer’s income
 Relationship between Purchasing pattern and region
 Relationship between Purchasing pattern and promotional mails
 Relationship between Purchasing pattern and Preference of online portal for shopping.

18
6.2.FLOW CHART:

DATA FLOW
SALES
DATABASE

R - STUDIO

Total sales prediction

Linear Regression
Analyzed using
Total quantity prediction

Products frequently
purchased together Frequency Pattern Growth
Analyzed using

Relationship between
income and buying behavior

K – Means Clustering
Relationship between Analyzed using
gender and buying behavior.

19
Data Visualization
6.3.Database Design:

Database design simply means the structure of the data that we hold for the purpose of
research. It is important for every research scholar to choose apt details from the dataset.

Attributes Name:

FIELD NAME DATA TYPE DESCRIPTION

Bill number Character and number Bill number of each

customer.

Unit Cost Numeric Unit Cost of individual

customers’ purchases.

Total Quantity Numeric Total Quantity of individual

customers’ purchases.

Total Sales Numeric Total sales of individual

customers.

Gender Character Gender of the customers.

Age Numeric 1 – Male

2 – Female

Region Numeric 1 – City

2 – Rural

Marital status Numeric 1 – Married

2 – Single

20
 Owning of car Numeric 1 – Yes
2 – No

Online Purchase Numeric 1 – Yes

2 – No

Average Spent Numeric 1 – Below 15,000

2 – Above 15,000/Week

Weekly Purchase Numeric 1 – Yes 2-No

7.SYSTEM TESTING AND IMPLEMENTATION

7.1.TESTING FUNDAMENTALS:

Simple linear regression:

The simple linear regression is used to predict a quantitative outcome y on the basis
of one single predictor variable x. The goal is to build a mathematical model (or formula) that
defines y as a function of the x variable. Once, we built a statistically significant model, it’s
possible to use it for predicting future outcome on the basis of new x values.) This is used to
analyse the sales data. Every departmental store retailer’s main aim is to have a high profit.
For that purpose and for the ease of purchasing for the customer, many retailers extend their
business to online promotional mails, offers, weekly/monthly basis of purchase etc., But not
every retailer will be successful in that. It is important to know some things which customer
use to purchase and to know about the buying pattern of the customer to increase the profit of
the Departmental store.
The mathematical formula of the linear regression can be written as y = b0 + b1*x + e,
where:
 b0 and b1 are known as the regression beta coefficients or parameters:
o b0 is the intercept of the regression line; that is the predicted value when x = 0.
o b1 is the slope of the regression line.
 e is the error term (also known as the residual errors), the part of y that can be explained
by the regression model.

Frequency pattern growth:

21
 The FP-Growth Algorithm is an alternative way to find frequent item sets without
using candidate generations, thus improving performance. For so much it uses a divide-and-
conquer strategy. In simple words, this algorithm works as follows: first it compresses the
input database creating an FP-tree instance to represent frequent items. After this first step it
divides the compressed database into a set of conditional databases, each one associated with
one frequent pattern. Finally, each such database is mined separately. Using this strategy, the
FP-Growth reduces the search costs looking for short patterns recursively and then
concatenating them in the long frequent patterns, offering good selectivity.

FP-Tree structure

The frequent-pattern tree (FP-tree) is a compact structure that stores quantitative

information about frequent patterns in a database .

Han defines the FP-tree as the tree structure as given below:

1. One root labeled as “null” with a set of item-prefix sub-trees as children, and a
frequent-item-header table.
2. Each node in the item-prefix sub-tree consists of three fields:

 Item-name: registers which item is represented by the node;

 Count: the number of transactions represented by the portion of the path reaching the
node.
 Node-link: links to the next node in the FP-tree carrying the same item-name, or null
if there is none.

K - Means cluster analysis:

Clustering is a broad set of techniques for finding subgroups of observations within a

data set. When we cluster observations, we want observations in the same group to be similar
and observations in different groups to be dissimilar. It allows us to identify which
observations are alike, and potentially categorize them therein. K-means clustering is the
simplest and the most commonly used clustering method for splitting a dataset into a set of k
groups.

22
Important steps in K –Means cluster analysis:

 Replication Requirements
 Data Preparation
 Clustering Distance Measures
 K-Means Clustering
 Determining Optimal Clusters

7.2.IMPLEMENTATION:

 To identify the total sales in order to improvise the profit of the store.
 To identify the total quantity in order to give consumers a satisfaction on their
purchase of a particular brand.
 To identify the products that are frequently purchased together in order to provide
combo offers to the customers.
 To segment the customers based upon the factor of preference of weekly
purchases.
 To identify which factors among owning of car, number of children, region,
income, marital status, preference of online shopping, promotional mails have
more influence on preference of weekly purchases.

23
 8.SCREEN LAYOUT AND SAMPLE CODE

SAMPLE LAYOUT:

Sales Prediction:

Fig:8.1. Sales Prediction using linear model

24
 Fig:8.2. Sales Prediction using linear model

DESCRIPTION:

Sales are the backbone of retail industry. It is important for every retailer to accurately
predict the total sales for gaining more profit. Here, linear regression is used to predict total
sales. Linear regression is a common Statistical Data Analysis technique. It is used
to determine the extent to which there is a linear relationship between a dependent variable
and one or more independent variables. Total sales are predicted using two independent
variables namely total cost and total quantity. According to this data, total sales act as
dependent variable and total cost and quantity act as independent variable. The regression
interception value is identified as -236.24873. This value is independent of Total quantity
and total cost. The predicted value of total cost and quantity are 1.269 and 55.286
respectively. This model based regression is constructed based on the following formula:

Total Sales = (-236.24873) + (1.269) * TotalCost + (55.286) * TotalQuant

25
Quantity Prediction:

Fig:8.3. Total Quantity prediction

26
 Fig:8.4. Total Quantity prediction

DESCRIPTION:

Despite trying to make customers loyal to their store, it is important to give them
quality products in the correct quantity. Quantity is as important factor as the quality of the
product. If the product’s quantity does not satisfy the customers, they may not be willing to
purchase the product. Sometimes this factor plays a major role in change of brand by the
customers. Here, linear regression is used to predict total quantity. Total quantity is predicted
using two independent variables namely total cost and total sales. Now, total quantity acts as
dependent variable. The intercept value is identified as 14.23 and it is independent of total
sales and total cost. The predicted value of total cost and sales are -0.0054 and 0.0045
respectively. This model based regression is constructed based on the following formula:

Total quantity = 14.23+Total cost*(-0.0054) + Total sales*(0.0045)

27
FP growth:

Fig:8.5. To identify the products that are frequently purchased together

28
DESCRIPTION:

This objective mainly focuses on identifying the products that are

frequently purchased together. This enables the shopkeepers to make some attractive combo
offers in order to make the purchase of customers a way much easier. This not only benefits
the customers but also the shopkeepers. Consumers may prefer different products of same
brand. At the same time, one particular product may not be sold properly. The shopkeepers
would include that product into the combo so that it would also get sold like the rest of the
products. Frequency pattern growth is used to analyze this. It compares the regular
expression patterns found in the values for the specified field and performs a frequency
analysis based on the patterns found. Here, The combination for the product “beer shampoo”
is checked and it has resulted 144 rules. The most frequently purchased products along with
“beer shampoo” are identified as Britannia cake, Britannia Marie gold biscuits, boost and
lever ayush tooth paste.

K-means clustering:

Fig:8.6.Segmentation of customers based on their preference for weekly purchase.

29
DESCRIPTION:

Income is the major factor that affects the purchases of the

customers. They are directly related to each other. K –means clustering is used to segment the
data according to their preference of weekly purchases. In this data, cluster 1 represents those
customers who are willing to purchase on weekly basis. 2 represents the customers those who
are not willing to purchase on weekly basis. From the above graph, it is understood that the
presence of customers who are not willing to purchase on weekly basis are more. This factor
influence the other factors such as income, owning of car, marital status, their wards,
preference of online shopping etc., So the promotional activity should be taken by taking into
account the factors that is mentioned above.

Fig:
8.7. Relationship between weekly purchases and promotional mails

Fig:8.8. Relationship between weekly purchases and promotional mails

30
DESCRIPTION:

Promotional activity is the life line of retail industry. These promotional activities
should be base upon the targeted customers. Here the relationship of promotional activities
and preference of weekly purchase is tested using chi-square. We test the hypothesis whether
the promotional activity is independent to the weekly purchases preference at the significant
level of 0.05. The probability value for this test of independence is 0.02178 which is less than
the significant value. So we reject the hypothesis. So, it is concluded that there is a
relationship between the promotional activities and weekly purchases. Likewise we test the
dependency of other factors like region, owning of car, number of children, marital status and
income have any influence on weekly purchases using ANOVA (Analysis of Variance). Here
it is understood that, the factors such as age, marital status, owning of car, promotional ails
and average spending have more influence on weekly shopping.
SAMPLE CODE:

Linear regression for predicting total sales:

31
data<-read.csv('D:/Project2020/7.2.20/salesBNGP.csv')

head(data)

tail(data)

print(data)

salesdata<-data[c(3,4,5)]

Testd<-salesdata[1:313,]

traind<-salesdata[171:313,]

Reg.Model<-lm(Totalsales~TotalCost+TotalQuant,data=traind)

traind<-salesdata[171:313,]

Reg.Model<-lm(Totalsales~TotalCost+TotalQuant,data=traind)

summary(Reg.Model)

newdata<-subset(data,Totalsales>=5000,select= c(billNogp,Totalsales,SalesPredicted))

newdata

newdata$billNogp<-reorder(newdata$billNogp,-newdata$Totalsales)

barplot(newdata$Totalsales,names.arg = newdata$billNogp,main = "sales before

prediction",xlab = "bill no.",ylab = "sales",las=2)

barplot(newdata$SalesPredicted,names.arg = newdata$billNogp,main="sales after

prediction",xlab = "bill no.",ylab = "sales predicted",las=2)

32
 data<-read.csv('D:/Project2020/7.2.20/salesBNGP.csv')

head(data)

tail(data)

print(data)

salesdata<-data[c(3,4,5)]

Testd<-salesdata[1:313,]

traind<-salesdata[171:313,]

Reg.Model<-lm(TotalQuant~TotalCost+Totalsales,data=traind)

summary(Reg.Model)

data<-read.csv('D:/Project2020/7.2.20/salesBNGP.csv')

newdata<-subset(data,TotalQuant>=35,select=
c(billNogp,TotalQuant,TotalQuantPredicted))

newdata

barplot(newdata$TotalQuant,names.arg=newdata$billNogp,xlab = "BILL NO.",ylab =

"QUANTITY",main = "QUANTITY BEFORE PREDICTION",col='purple',las=2)

barplot(newdata$TotalQuantPredicted,names.arg=newdata$billNogp,xlab = "BILL
NO.",ylab = "QUANTITY",main = "QUANTITY AFTER
PREDICTION",col='purple',las=2)

Linear regression for predicting total quantity:

Frequency Pattern Growth:

33
library(rCBA)

data<-read.csv('D:/Project2020/7.2.20/d_trans2.csv')

head(data)

data$SalesType=as.factor(data$SalesType)

dafa<-data[, c(11:20,174)]

head(dafa)

data.trans<-as(dafa,"transactions")

data.trans

data.rule<-rCBA::fpgrowth(data.trans,support=.2,confidence = 0.2,

consequent = "beer_shampoo")

data.rule

inspect(data.rule)

plot(data.rule)

library(arulesViz)

plot(data.rule)

34
K-means Clustering:

customer=read.csv("D:/Project2020/20.2.20/customer_profile.csv",head=TRUE)

head(customer)

cus.data=customer[251:550,2:12]

cl.kmodel<-kmeans(cus.data,2)

cl.kmodel

cus.clust<-data.frame(cus.data,cl.kmodel$cluster)

head(cus.clust)

35
 library(cluster)

clusplot(cus.clust, cl.kmodel$cluster, color=TRUE, shade=FALSE, labels=2,

lines=0)

Chi – Square:

tbl = table(cus.data$Pro_mails, cus.data$Weekly_Pur)

tbl

chisq.test(tbl)

ANOVA:

head(cus.data)

mod.a <- aov(Weekly_Pur~., data = cus.data)

summary(mod.a)

9.CONCLUSION

In this paper, R programming is used to analyze the sales data of a retail departmental
store using certain machine learning algorithms and clustering such as Linear model
regression, K – means clustering, and frequency pattern growth. The main intension of this
research is to aid the retail shop in identifying the accurate sales and quantity value,
identifying the products that are frequently purchased, segmenting the customers into two
clusters based upon their weekly purchases and relationship between gender and purchasing
behavior. This would improve the functioning of the store. Because by acquiring this result,
they could improve sales price, increase quantity and make customers loyal to their sales,
make attractive combo offers and customers can be segregated based upon their preference of
weekly purchases.

36
 REFERENCES:

[1] Adil Zia M., Khalid Azam, “Knowing Consumer Loyalty A comparative study of
Organized and Unorganized retail”.

[2] Dr Amit Mittal, “Store Attribute Salience- A comparison of grocery & apparel shopping
scenarios” ISSN 0974 – 9497, Volume 3, Issue 4/4, November 2009.

[3] Che Aniza Che Wel, Siti Rahayu Hussin, “Important Determinant of Consumers’ Retail
Selection Decision in Malaysia”, World Review of Business Research, Vol. 2. No. 2, Pp. 164
– 175, March 2012.

37
[4] Cherukuri Jayashankaraprasad, “Effect of Situational Factors on Store Format Choice
Behaviour in Food and Grocery Retailing in India - Multiple Discriminant Analysis”, IBSU
Scientific Journal, 4(2), 5-33, 2010.

[5] Dr. Harsh Sharma, “Influence of family size on retail shoppers’ perception of Indian
traditional grocery stores’ attributes”, International Journal of Multidisciplinary Research
Centre, ISSN: 2454-3659 (P), 2454-3861(E), October 2015.

[6] Indrila Goswami Varma, Ms. Rupa Agarwal, “Online Buying Behavior of Homemakers
in Western Suburbs of Mumbai and Social Media Influence”, IOSR Journal of Business and
Management (IOSR-JBM), p-ISSN: 2319-7668. Volume 16, Issue 8, August 2014.

[7] D.Manimegalai, “A Comparative Study on Consumer’s Behaviour towards Organized and

Unorganized Retailing in Central Chennai”, International Open Access Journal, ISSN No:
2456 – 6470.

[8] Mathura Valli C. R. “An investigation of consumer preference towards store brands
purchase in Madhurai district, Kalasalingam academy of research and education”, June 2013.

[9] Navneet Gera, Ritu Bajaj, “A comparative study on the prospects of label brands to that
of national brand in FMCG retail in Delhi”, I J A B E R, Vol. 14, No. 6, June 2016.

[10] Piyush Kumar Sinha, Arindam Banerjee, “Deciding Where to Buy: Store Choice
Behavior of Indian Shoppers”, Vol. 27, No. 2, April-June 2002.

[11] Prateek Kalia, Dr.Navdeep Kaur, “A Review of Factors Affecting Online Buying
Behavior”, Journal of Management and Technology, Vol. 11, Number 2, July 2016.

[12] Raghuram Naga Venkata J, Ravilochanan.P, “The Role of Supply Chain Management
on Organized and Unorganized Grocery Retailers-A Review on Indian Perspective”,
Canadian Center of Science and Education, Vol. 10, No. 22 2014.

38
[13] Rama Venkatachalam, Arwah Madan, “A Comparative Study of Customer Preferences
towards Fresh Groceries: Organized v/s Unorganized Retailers”, 2012.

[14] Shruthi G. Jogur, “Consumers’ perception towards shopping in Siddeshwar super

bazaar”, Kousali Institute of Management Studies.

[15] Sugandha Agarwal, “An empherical study of various factors influencing the behavior
of consumers towards fast food joints in Indian market.

BIBLIOGRAPHY:

39
40