Data Analytics for E-commerce Profit Enhancement: A Case Study of a Book Segment

This report investigates the application of data analytics to enhance the profitability of an e-commerce company specializing in books. Using a dataset of book sales, the report analyzes key factors influencing sales performance, including geographic regions, product popularity, and the potential impact of free shipping. Data mining techniques, including hierarchical clustering, are employed to identify areas for improvement. The report provides actionable recommendations for the company, including targeted marketing strategies, product prioritization, and regional focus. The findings demonstrate the value of data-driven insights in optimizing e-commerce operations and achieving sustainable growth.

ICT706

Executive Summary
Electronic commerce or e-commerce is a business model which helps an organization to do
business on an electronic network or commonly known as the internet. E-commerce has almost
all the product or services like books, music, financial services and plane tickets.
This report mainly covers how to cover those aspects which are going to give a hike to the
performance of the company. And those aspects are going to be stated by the data scientist to
ensure success in the performance of the company.
All the useful information about the company is going to be collected through a SWOT analysis
method. To know about the current business operations and the solution provided by the
company in more depth qualitative data analysis method will be taken in use.
To increase the performance of the company it is necessary to have an information architecture
which is customer centric which is productive. The content and SEO should be integrated to
bring your business in front.
The product should be advertised in such a way that it becomes easy for the consumer to search
the product on the e-commerce sites. There should also be a relaxation for the consumer on the
shipping charges of the product to improve the business of the organization.
1

Table of Contents
Executive Summary.........................................................................................................................1
List of Assumptions and Abbreviations..........................................................................................3
Background......................................................................................................................................4
Introduction......................................................................................................................................5
Research Methodology....................................................................................................................6
Descriptive or Analytical:............................................................................................................6
Applied or Fundamental:.............................................................................................................6
Quantitative or Qualitative:..........................................................................................................6
Conceptual or Empirical:.............................................................................................................7
Analytical Findings..........................................................................................................................8
Dataset..........................................................................................................................................8
Data Mining:................................................................................................................................9
Divisive method:........................................................................................................................10
Agglomerative method:..............................................................................................................11
Recommendation for the Company:..............................................................................................14
Implementation plan:.....................................................................................................................15
Code Implementation.....................................................................................................................16
Conclusion:....................................................................................................................................18
References:....................................................................................................................................19
Appendix........................................................................................................................................20
List of Figures
Figure 1: Books Dataset Snippet.....................................................................................................8
Figure 2: Training the Data using the Linear regression.................................................................9
Figure 3: Testing the data using the Linear Regression................................................................10
Figure 4: Sales Graph for Different Region..................................................................................11
Figure 5: Region Recommendation...............................................................................................12
Figure 6: Product Recommendation..............................................................................................12
Figure 7: Code for Data Prediction................................................................................................16
Figure 8: Final Prediction Over Dataset........................................................................................17
2

List of Assumptions and Abbreviations
The Abbreviations that are used in this report are:
 Comma Separated Value – CSV
 Tab Separated Value – TSV
Assumptions for this report:
 A test dataset is developed for this process
 The Test Dataset will be divided into the Training Set and the Test Set
 Geographic Locations is added with it
 The Products are unique in this
 The need for a Delivery Person
3

Background
This E-commerce company has been set up a long time back. At the beginning as it was not very
recognized it used to deals in very fewer regions and had a less amount of products to start with
but after some time passed it expanded its business from very few regions to a lot of regions and
not only with regions but it also expanded in the products because of which the company start
getting recognition .
This company works on almost all the domains which include Gadgets, Books, Toys, Clothes
and household items. Thinking about its customers the company delivers the product on the
doorsteps of the customer.
After having a successful glorious year for so long the company has hit the rocks and is an
immediate need for the strategies to expand its business like before or more than that. And to do
that the company has thought of taking in use the data analytics to get back its position in the
market. Through data analytics, the company will know about the regions from where they can
attain the highest performance.
4

Introduction
As the Current scenario is where everything revolves around the e-commerce all the companies
are just competing with each other by surpassing other to be on the top but since the requirement
of the market changes every time. It is not very easy to cope up with the situations.
In this report, it mainly highlights an e-commerce organization who is searching for methods to
enhance the profit of the company with the help of data analytics. Its main aim is to know the
estimate of the monthly sales data for the segments of the book of the company. This report
stores the details about the products which has been sold in the books segment by the company
in the last few months.
The main agenda of this report is to work on the possibilities which are going to increase the
profit of the company by taking help of the Python programming to work on that possibility.
This report also contains the method which is going to produce the output.
5

Research Methodology
Research is a process which defines and redefines all the problem that came up, working on
assumptions or answers, gather and evaluate the data and reach to a conclusion and at last, all
that endings are tested in order to see if it just the assumptions or not. Research is putting extra
knowledge to already available knowledge to make it more advanced. Research is something
where to get to know more about something to actually available data by going in more deep
through study, observation, comparison, and experiment.
Research has been categorized as follows:
Descriptive or Analytical:
In descriptive research, it includes surveys and the inquiries of all the found facts. The main
thing which descriptive research revolves around is on what has happened and what is
happening. In descriptive the things which are measured is the frequency of shopping,
preferences of people or something like that. On the other hand, analytical research deals with
the information which is already available and through that make the critical evaluation.
Applied or Fundamental:
Applied research is something which works on giving quick solutions to the problem of an
organization. Fundamental research deals with the thought of a theory.
Quantitative or Qualitative:
Quantitative research where quantity is measured. It works for something related to quantity or
amount. Qualitative research revolves around the things related to quality which includes
behavior science where it researches about the human behavior. To apply qualitative research is
comparatively difficult and it needs guidance from experimental psychologists
6

Conceptual or Empirical:
Conceptual research deals with some conceptual idea. Philosophers take help of these for
generating ideas and concept or make some change in something in the ideas which already
exists. Empirical research depends on experience or observation. It is based on data where the
conclusions are verified by observation or experiment.
7

Analytical Findings
The report uses a dataset of books segment. The dataset includes only the books which are sold
by the company.
Dataset
The dataset includes:
 Product name,
 Product price,
 Shipping type
 Monthly sales
 Geographic region
 No. Of customers who bought the product,
 Customer type
Figure 1: Books Dataset Snippet
Figure 1 shown above illustrates a part of the dataset which also includes the metadata in the
books dataset
8

Data Mining:
Data mining is the method which companies take in use where all the raw data is converted into
useful information. With the help of software, patterns can be looked in large batches of data, it
will help to learn more about the customers and develop marketing strategies which are effective
and help in increasing the sale and decreasing the cost.
As the process of data mining helps in determining the sales and give high revenues it is used by
a lot of e-commerce companies. Data mining is dependent on data collection and storing that
data in warehousing. Its main aim is to gather the information from the dataset and deliver it in
such a way that it is understandable. For our company data mining can help in by providing the
ways to improve the productivity, what is going to happen if free shipping methodology will be
put in action, which region should be targeted to increase the sales, which product should be
given the highest priority to improve the sales.
9

Figure 2: Training the Data using the Linear regression
10

Figure 3: Testing the data using the Linear Regression
The output of the data mining is according to the data type. Clustering contains the dataset
having the same kind of data. Hierarchical Clustering deals with forming clusters that have a
predetermined ordering from top to bottom. It is categorized in two methods:
Divisive method:
This method is also considered as top-down where every observation is assigned to an individual
cluster and then divide the previously achieved cluster into two alike clusters. In comparison
with agglomerative method gives more accurate hierarchies but when the concept is considered it
is more complex.
11

Agglomerative method:
Each observation is assigned its own cluster and find the similarity in each cluster and combine
the two same clusters. This method is also called bottom-up clustering.
Figure 4: Sales Graph for Different Region
The geographic regions to target:
The geographic regions that should need to be worked upon are:
 Brisbane
 Boulis
 Hamilton
 Roma
 Bridport
 Taree
The regions mentioned above Hamilton is the place which needs to be worked upon the most
observed from the figure given above. If the company is keen on increasing its sales then these
are the geographic locations where the main focus should be on by keeping Hamilton on the top.
12

Output:
Figure 5: Region Recommendation
Brisbane should be focused on the sales of books and particularly for ‘A kiss in Time’.
The product which needs to be prioritized:
Figure 6: Product Recommendation
‘Up from the Grave’ needs to be Prioritized for the sales.
Free Shipping Impact On Product Sale:
Giving your customers their products on their doorsteps that too for free is something which
every customer seeks for. Implementing this will help the organization in increasing its customer
which is going to balance the shipping charges if the cost of the product is also decreased then it
needs to make sure that it does not decrease the quality of the product to maintain the reputation
of the company. Since other companies charge some extra money for shipping it is going to
favor our company more and will attract more clients.
To improve profit of company:
To improve the profit of company the geographic regions needs to be worked on. Some policies
need to be implemented in those geographic regions that are where the companies are lacking.
Those products which are not being sold should not be stored for long and whenever those
13

products are ordered needs to be bought at that time only. Digital marketing should be done in
those regions which have fewer clients. Those products which are comparatively bought less for
those free shipping should be available.
14

Recommendation for the Company:
The sales will be increased if sales happen in festive seasons. Products should be delivered for
free, If the customer is unsatisfied then there should be a policy of returning the product. The
product ordered by the customer should reach to the customer on time. If the website is not that
attractive, then its user interface should be improved. Other recommendations include:
 A training Team
 Better Advertising System
 Scalable Data Storage
 Better Data Cleaning
15

Implementation plan:
The regions which are lacking behind in sales needs to be improved by implementing some plans
so that it can increase some client in that region. Which is going to help in the overall growth of
the company. Then the products which have been bought the least some marketing strategy
should be implied to increase the sale of those products.
16

Code Implementation
#Library Import
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.cross_validation import train_test_split
from sklearn.preprocessing import PolynomialFeatures
# This Will Import The Dataset of the Books
getDataset = pd.read_csv('DatasetBooks.csv')
xData = getDataset.iloc[:,2:3].values
yData = getDataset.iloc[:,3:4].values
# This will split the dataset into two separate sets Training Set and the Test Set using train_test_split
library in the Sklearn
xDataTrain, xDataTest, yDataTrain, yDataTest = train_test_split(xData, yData, test_size = 1/3.0,
random_state = 0)
#Start Linear Regressison
linearmodel=LinearRegression() #create a variable for the Linear Regression from the Sklearn module
linearmodel.fit(xDataTrain , yDataTrain) #fitting data accordingly to the Linear Regression
predOnY=linearmodel.predict(xDataTest)
plt.scatter(xDataTrain , yDataTrain, color='red') # Scatter Plot
plt.plot(xDataTrain, linearmodel.predict(xDataTrain), color='blue') # Plotting Traing Data and the
prediction over it
17

Figure 7: Code for Data Prediction
Figure 8: Final Prediction Over Dataset
Figure 8 will show how the company sales are going to be according to the Books price. This
model shows how will the test and training data were helpful in creating a final predictive model
for the whole dataset.
18

Conclusion:
This report includes the methods which are going to help in increasing the sale of the company in
those regions where the company is lacking behind. In this report, a dataset has been given of a
product segment of an e-commerce company. This is for the purpose of analyzing and
increasing the profit of the company. It mainly revolves around the regions and products which
needs work to improve the sales of the company.
A specific data mining has been applied to the dataset and then clustering has been done which
helps in which products need the attention to attain more profit. Hamilton is the region which
needs the most attention to increase the sale of the company.
An Implementation plan has been given as a data analyst about the factors which need the
attention of the company and helping the company in attaining the profit.
19

References:
 Fontinelle, A. (2018). Electronic Commerce (e-commerce). [online] Investopedia.
Available at: https://www.investopedia.com/terms/e/ecommerce.asp [Accessed 31 May
2018].
 Heirarchial clustering (2018). Hierarchical Clustering. [online] Saedsayad.com. Available
at: http://www.saedsayad.com/clustering_hierarchical.htm [Accessed 31 May 2018].
 Lyngbø, T. and Lyngbø, T. (2017). E-commerce companies can improve revenue by
following these 5 tips - Search Engine Land. [online] Search Engine Land. Available at:
https://searchengineland.com/e-commerce-companies-can-improve-revenue-following-5-
tips-271547 [Accessed 31 May 2018].
 Modares (2004). [online] Modares.ac.ir. Available at:
http://www.modares.ac.ir/uploads/Agr.Oth.Lib.17.pdf [Accessed 31 May 2018].
 Staff, I. (2018). Data Mining. [online] Investopedia. Available at:
https://www.investopedia.com/terms/d/datamining.asp [Accessed 31 May 2018].
20

Appendix
21