E-commerce Data Analysis: Sales Trends and Strategic Improvement

This report provides an executive summary of e-commerce principles and their importance in online business. It details how data scientists can improve company performance using SWOT analysis and qualitative data analysis. The report uses Return on Investment to measure efficiency and suggests solutions like free shipping and targeted marketing strategies. It covers the company's background, its expansion, and the increasing competition in the e-commerce market. Data analytics is proposed to address declining sales, focusing on profitable locations. The research methodology includes fundamental, applied, normal, revolutionary, quantitative, and qualitative research types. Analytical findings are based on a Books segment dataset, employing data mining and clustering techniques to predict monthly sales and identify key geographic regions. Recommendations are made for product prioritization and assessing the impact of free shipping. The implementation plan aims to enhance company performance and achieve target profits through data-driven strategies.

Executive Summary
E-commerce is a word which deals with the business of any type which revolves around
transferring the information on the internet. It deals with a variation of business which includes
the services within the company (Networksolutions, 2014). Nowadays it is considered as one of
the important stuff of the internet. E-commerce helps the consumers in exchanging goods
electronically without to have worry about the time and distance. To have a successful online
store there must be a knowledge of E-commerce principles and how it is supposed to be handled
to do the online business. The guidelines need to be properly understood to have a successful
online store.
This report is making sure to work on those technological measures which are going to help the
company in improving their performance. Data Scientist will be providing solutions to make sure
that the company receives success in all aspects.
All the information which is used will be gathered using SWOT analysis method. There is also a
method stated as qualitative data analysis which will be studied in deep dealing with the present
operations of the business in the company and all the solutions possible.
Return on Investment is the method to measure the performance and helps in calculating the
efficiency of the company. Which can be considered as one of the effective measures for the
company.
There are many solutions to enhance the performance of the company which includes the option
of free shipping. There should be a marketing strategy to make sure that the product is advertised
in such a way that it becomes simple to find it on e-commerce sites.

Table of Contents
Executive Summary.....................................................................................................................................1
Background..................................................................................................................................................3
Introduction.................................................................................................................................................4
Research Methodology................................................................................................................................5
Types Of Research...................................................................................................................................5
Fundamental or Basic Research:..........................................................................................................5
Applied Research:................................................................................................................................5
Normal and Revolutionary :.................................................................................................................5
Quantitative and Qualitative Research :..............................................................................................5
Other Type Of Research :.....................................................................................................................6
Analytical Findings.......................................................................................................................................7
Dataset....................................................................................................................................................7
Data Mining:............................................................................................................................................7
Clustering:...............................................................................................................................................8
Predicting Monthly Sales:........................................................................................................................8
Geographic Regions:................................................................................................................................9
Recommendations:....................................................................................................................................12
Implementation Plan:................................................................................................................................12
Conclusion.................................................................................................................................................13
References.................................................................................................................................................14
Appendix:..................................................................................................................................................15

Background
In the beginning when this company was just started used to deliver their products in a very
limited region as it was not that big of a company in the start. As it was a small company it used
to have not a very vast amount of products but after the company started expanding it started
flourishing in different regions because of which the company emerged as a one of the popular
company.
This company works on a lot of domains which includes Gadgets, Books, Books, Clothes and household
items. This company has provided its customer with the facility of delivering the products to the customer
on their location which they have provided.
As in the recent few years, a lot of e-commerce company has been set up which results in a lot of
competition with some new companies coming they also bring some new strategy to attract the
consumers. As a result of all of this, the performance of the company has been keeping on decreasing as
the company has started losing its clients.
As mentioned above due to the development of new e-commerce companies the company is facing a lot
of degradation in its consumers which is resulting the company is being in the loss which is affecting the
overall impact of the company in the market. So, to overcome all these situations the company has
decided on using data analytics to gain perspective on the sales. Data analytics is going to help the
company in putting attention to those locations to have some profit in sales.

Introduction
Nowadays when the market is all about e-commerce a lot of companies are battling with each
other to just have that position on the top , to gain the most profit but as the market keeps on
changing in a blink of an eye it is not very easy to cope up with the situations making it harder
for every company to achieve their goal with the old tactics.
This report is going to provide the ways to the company to have that target profit of theirs with
taking help from data analytics. This report is going to deal with assuming the sales data of each
month of the company for the Books segments.
There are details of all the products which has been sold by the company in last few months in
the sector of Books segments. The main goal of this report is to go through the sales of the
company in last few months and to get to know about how the company can have profit taking
use of python programming for several tasks. Naïve Bayes helps in making the production more
strong which is done in excel. This report also has the methods for generating the output.

Research Methodology
In Research methodology problems are being solved through an organized way.It helps in
carrying out research. It is a procedure through where the researchers explain the work they
need to elaborate and assume the phenomena is called the research methodology. It is known as
study of methods also which helps in gaining knowledge.
Types Of Research
Fundamental or Basic Research:
Basic research is inspecting the basic principles of a specified action to occur(Arxiv,2011). It is a
considered as theoretical research also.In basic research the natural happenings and basic
research is studied. Basic research provides a well organized and detailed idea of a problem and
makes easier the conclusion on logical explanation.
Applied Research:
In Applied Research it does not have the objective to generalize it deals with individual cases.
Applied research focuses on any variable which is going to make a difference which is desired. It
tries to tell how things can be changed. All the problematic facts are cleared.
Normal and Revolutionary :
Basic and applied research are of two kinds which are: normal research and revolutionary
research. Normal research works on a set of rules considered as a paradigm. Normal research is
similar to solving a puzzle, solutions are found but there is no change in rules.
Quantitative and Qualitative Research :
In Quantitative research it deals with measuring the quantity. The output which research is going
to give is a collection of numbers. Qualitative research majorly deals with quality which is not
numeric, needs description, requires reasoning and deals with words.

Other Type Of Research :
Comparative research, explanatory research, Action research, exploratory research and
comparative research are another type of research.

Analytical Findings
In this report, a dataset has been taken in use which is based on a Books segment. So, the dataset
is going to have the entries of Books which is going to be sold by the company.
Dataset
The dataset of the Books segments include following:
 Product Name,
 Product Price,
 Monthly Sales,
 Shipping Type,
 Geographic Region,
 Customer who bought the product, and
 Customer Type
Figure 1: Dataset
Data Mining:
Data Analysts take use of data mining to find some patterns which are useful in a large amount
of data and then use that data in a useful manner which a businessman can easily understand.
Data mining is dependent on the feasible collection of data and then store that data in
warehousing(SearchSQLServer,2018) . For prediction analysis most of the

Companies use the process of data mining to estimate their sales and then have a high revenue.
Data preprocessing is done before data mining to make sure that no values are missing, reduction
of noise and clean the data. Cleaning of data is done by python code. The main purpose of this
process is to take the information which is useful and present in such a way that it is easy to
understand. Data Mining has been performed in this report for which geographic regions need to
be targeted, what are the products which need to get priority for sales, What if free shipping will
start happening how it is going to impact the company and how productivity can be improved.
Clustering:
After the process of data mining has been implemented the output received from that process is
going to be clustered according to the type of the data and python program helps in analyzing. In
clustering, hierarchical clustering has been chosen as the methodology to do this which is
considered as the method for analyzing the cluster through hierarchical models of the
dataset(Epubs.siam,2018). There are also two approaches in hierarchical clustering which
includes Divisive approach also considered as top-down approach wherein this method the
individual cluster has been assigned an observation and the cluster which has been previously
achieved has been distributed into two clusters who are alike. This gives more accurate
hierarchies when the methodology is compared with the agglomerative approach but if the
concept is taken into consideration then it is more complicated than agglomerative approach.
On the other hand in agglomerative approach, every observation has been assigned its own
cluster and to detect the similar traits in each of the clusters and then have the combination of
both the same cluster. The agglomerative method is also considered as bottom-up clustering. It is
comparatively simpler in terms of concept with the divisive approach.
The relation between the attributes has been shown in the figure given below taking help of
hierarchical clustering. And the other figure given below shows that how much distance is there
between the attributes.
Predicting Monthly Sales:
Naïve Bayesian has been derived from the Bayes theorem having independent assumptions
between the predictions. It is not that difficult to build as it does not have any complex repeating
parameter to estimate which clearly indicate to one thing that it is very useful for a large number
of datasets. Although it is simple its performance is amazingly well and is taken to use a lot as it
is considered better than classification methods(Saedsayad,2018). It is an algorithm which gives
the permission to the data analysts to implement rules on a set of observation. To take help of it
firstly the classifier is trained on the dataset which is testing which signifies what set of inputs
are needed to be expected. The algorithm finds out the rules of classification from the dataset
which is training taken already in use that is going to be taken again in use for predicting the
observation on the dataset which is predicted.

Geographic Regions:
The geographic that does not have a large number of clients and that needs to be worked upon
are:
 Hurtsville
 Logan Hills
 Sydney
 Mildura
 Booie
 Cairns
 Officer
 Brisbane
The regions mentioned above demonstrates the geographic locations where the company really
need to work upon to improve the sales and out of all this, the least sales has been seen in
Sydney where the company has the lowest sales. So, the company really needs to work the best
in these locations and for Officer, they need to bring their A-game to gain its previous popularity
and to be on the top.
A product that needs to be prioritized :
 Max's Breakfast (Max and Ruby),
 Too Small for My Big Bed: Sleep Tight in Your Own Bed Tonight!,
 National Geographic Kids Sharks Sticker Activity Book: Over 1,000 Stickers! (NG
Sticker Activity Books),
 Franklin Has a Sleepover
 Lon Po Po: A Red-Riding Hood Story from China
 The Pout-Pout Fish book and CD storytime set
The above books are some of those books that need to be prioritized above all other books.
Among them, the book that should be prioritised is “The Pout-Pout Fish book and CD storytime
set” for the better sales increase.
Figure 2: Recommendation Based on analysis
Impact on sales after free shipping:
The company performance has degraded a lot to boost up their performance and to make sure
that they achieve their goals they need to bring some marketing strategies for their product and
free shipping of the product is one of them. Customers not only love that their product is being
delivered to them for free but there isn’t any shipping charge of bringing to them their product is

like a cherry on the top. So, if the company really wants to increase their sales they need to apply
this marketing strategy to bring back their position.
Methods to improve profit for the company:
To improve the profit of the company the company needs to work upon the geographic locations
that have been facing the most degradation or in the other words that location who is not gaining
any profit. Those places needs some work to improve the performance of the company there so
such marketing strategies need to be planned that is going to increase the sales on those areas
where the company is facing loss not only the geographic locations but the company also need to
work upon the products which are not being sold as much they should. Those products should
become the priority and they should be sold off.
Figure 3: Result of Training Set over the Books Data

Figure 4: Training Set Result for the Books Dataset

Recommendations:
The strategies which the company should work upon to improve the sales are that there should be
sale of the products provided by the company in a month or two so that it can attract a lot of
customers , the products which are being delievered there should be no shipping charge on that
product, if the customer is not finding the product as they thought of it when they bought it then
in that case there sould be a return policy for the customers and the appearance of the website
should be such that it makes the customer feel to go and buy the product.
The Company could further change their Data Storage Plan along with the maintenance of the
warehouse. Further, the company needs to be well aware in implementation for the stock and
keep updating it every once in a while.
Implementation Plan:
The implementation plan includes the digital marketing which is taken in use should be fine. The
company should put its attention on facing those regions which are deteriorating in the market as
if these regions had an increase it is going to affect the company all over in a positive way. And
the other focus of the company should be in selling those products which are not being sold off.
The Implemnatation paln is based over the analysis that was done in order to predict the nature
of te data.