Last few generations were trying to make their day to day activities very simple. They mostly use the internet to accomplish their target. E-commerce is one of the most popular buying and selling system used by millions of people around the world. This project is a proof of concept to
Table of Contents
List of Abbreviations
UBA – User Behavior Analysis
E-Commerce – Electronic Commerce
SEO – Search Engine Optimization
GA – Google Analytics
UML – Unified Modelling Language
OAP – Online Analytical Processing
OLAP – On-Line Analytical Processing
RDBMS – Relational Database Management System
DBMS – Database Management System
JVM – Java Virtual Machine
FDA – Flash Data Archive
XML – Extensible Markup Language
OLTP – Online Traction Processing
E-commerce sites are referring to the use of an electronic medium to carry out commercial transactions. Most of the time it focuses to sell products online. But E-commerce is not limited to that, it also covered followings,
- Real-time management of product availability (Stock or Inventory) –
Real time inventory availability facilitates buyers and company employees to see the available stock of a product. Real time means the management system updates inventory levels immediately after the inventory is sold
- Online payment –
The online payment system is a way of paying for a goods or services electronically, instead of making payment with cash or a check, in person or by mail. Online payments will play major role in today’s E-Commerce sites. Major reason of online payments is, it gives the ability for buyers pay easily and securely purchase products from anywhere in the world. Also, this payment method help sellers to handle their transactions in a secure, trusted path. Major online payment systems are PayPal, Amazon payments, Google checkout, etc. Other than that users can also pay using Credit card, Debit card, Smart card, E-Money, Electronic Fund Transfer (EFT).
- Package Delivery tracking –
Package tracking or logging is that the method of positioning shipping containers, parcel post at varying points of time during storing and package delivering. This feature most important to users because it allows users to view complete delivery details and users can estimate the package delivery date. In e-commerce the tracking details are provided by some of the sellers. There are two types of methods to get the tracking status provided by e-commerce companies. Those are by user login or by typing order number. The top players of e-commerce, such as Amazon, Staples, Walmart provide customers to sign in to their account or input order information to get the latest delivery status.
(“5 e-Commerce Best Practices for Package Tracking,” n.d.)
- After sales service –
A collective term for attention to a transaction that usually takes place when the sale (and often after delivery) is accomplished. Normally after sale service functions are, installation, training, part replacement support and warranty-related repair. In e-commerce after sales services are, refunding, give another new item, item replacing etc. After-sale service is an one of the most important feature in e-commerce because of the normally greater distances between sellers and buyers in international trade.
(“Ecommerce:Tips to improve the after-sales service | Mukund Chaudhary,” n.d.)
- Preparation of estimates online
- Supply virtual shopping cart –
An electronic commerce tool (software or service) that is the user-interface for the customer to shop at online stores. It allows users to place items in to their own “shopping basket”, and the cart keeps user selected items for a programmed length of time. Extra options such as size options, different colours, quantity of order item, and specific item links can be found linked from the shopping cart. Taxes and shipping costs can also be automatically calculate from inside the shopping cart, once a shopper inputs his or her shipping address. The virtual shopping cart also provides important information for the dealer, which is often transparent to the shopper including a cart number to track the order, and a cookie to provide with some limited tracking details about customer.
- Online auctions –
Virtual auctions on the internet. The seller sells the product or service to the person who bids the highest price. For sellers, online auctions open up new sales channels for new products and offer buyers favourable purchasing conditions.
E-commerce sites became a more popular among all younger and older generations around the globe, because it made a bridge on trading between countries, cities, villages in the world that make the world trading like nothing before. When we consider about most popular e-commerce sites such as mydeal.lk, wow.lk, they have thousands of sellers and millions of active users. As well as these sites make thousands of transactions per day and thousands of items adding daily. But it is not possible to show each and every items for each users. E-commerce looks like a social network on trading category among buyers and sellers. If we can make a knowledge base of buyers activity, that would be great to display related items for specific single user.
The aim of this project is to develop a proof of concept e-commerce site that can describe how user behaviour analysis (UBA) can be applied to an e-commerce web application to dynamically rearrange the content of the site for specific users. Dynamically rearrange mean, when users purchase any item or search for an item system will store the related data for the recent activity. After that system will identify other related items that match to recent purchased item. Then it will show the matched items on specific user’s home page.
This system will rearranging original site items list according to the user actions history on the e-commerce site. Another purposes of analysing the user behaviour on the e-commerce site is increase revenue by suggesting more deal that customer really looking for. As well as we can turn browsers in to buyers and build customer loyalty.
Nowadays buying and selling items through internet is very popular among every generation around Srilanka. Each year the number of e-commerce site users are increasing than normal retail shopping, because it is time saving and easy.
Most of e-commerce sites such as mydeal.lk, wow.lk are displaying their deal items according to recently updated items first or randomly selected items on their home page. Also some are showing deals under categories such as Home and Kitchen, Computers and accessories etc. because of that every user see the site with same content time to time. If user needs to find an item he/she must search with a specific keyword related to the item or scrolling down to find the item.
That will not an effective way for interact customers with the site and motivate customers to purchase items, because users will not interest with items that will not related to them. This will reduce the purchasing rate of the site and also decrease active users of the site.
To develop an e-commerce site that analyse user behaviour on the site such as purchasing history, browsing history etc. Then rearranging items on the site according to the behaviour of the buyers to interact users with the site, increase daily visiting count of the site, increase revenues of sellers and increase purchasing rate of the site.
1.3. Scope of the Project
The main scope of the system would be to suggest suitable items according to the user behaviour and increase purchasing count of the site. Other scope areas will be:
- Suggest offers that is suitable for user
- Optimize categories according to the user need
- Adjust items under each categories that is best match for user
Developing a software application with users sensitive data need to be more concerned, which require more care when using them. Well established and actively maintained e-commerce platform is that many of these products have done extensive testing and compliance verification due to the potentially sensitive nature of the information they work with and services they provide. The platform development should be active providing updates regularly, especially soon after major bugs or vulnerabilities are discovered. As the e-commerce becomes increasingly dynamic, the traditional static modelling may not be able to deal with it. One solution is to use agile modelling that is characterized with flexibility and adaptability. On the opposite hand, data mining applications need larger diversity of technology, business skills, and knowledge than the traditional applications. That means it will be benefit a lot from features of agile software development. (“Learn 4 Easy Ways to Protect Your Customers in an Online Store,” n.d.)
- Project Deliverables
- Design of the proposed web application
- Implementation of the web application with following features
- Suggest best matching deals and offers according to the user
- Optimize suitable categories according to the user
1.7. Work Breakdown Structure
Chapter 2: Literature Review
This Chapter describes the literature review of the primary research and the secondary research carried out by the author in order to identify the problems and challenges with the existing e-commerce system and the technologies used by the author to develop the e-commerce web application.
2.1. Primary Research
Primary research is mainly consider about internet usage on e-commerce, e-commerce market in Sri Lanka and existing e-commerce sites in Sri Lanka.
2.1.1. Internet Usage on e-commerce in Sri Lanka
Internet is the main source to connect with e-commerce sites. By using e-commerce site users will able to find products that they looking for and place an order, make an payment with in few minutes without any hastle. It save lots of money and time for buying a product rather than using typical methods like buying item from super market.
Since the year of 2016, there are cellular mobile connections over 1.5 million and over 0.3 million broadband and dial-up internet connections subscribe by Sri Lankans, increasing Sri Lanka’s internet saturation to a thirty percent and to total internet users to a 6.1 million.
Although Sri Lankan government’s recent removal of telecommunication services from the fifteen percent value-added tax exempt list and the steps taken by 2017 budget to remove tax holidays given to data leading to 50% effective taxation on mobile services, the Sri Lankan’s urge to increase their connectivity for educational, recreation and occupational reasons have remain sturdy through the years.
Sri Lanka’s increased internet connectivity also have given a boost to the Sri Lankan’s presence in the social media, especially on the Facebook, the favorite local online hangout. Sri Lanka’s internet connectivity is improving at a steady pace.
Figure 1 : Internet connectivity of Sri Lanka by 2017
Introduction of the Electronic Transactions Act, No.19 in 2006, which facilitates e-commerce by removing legal barriers and ensuring legal certainty both within Sri Lanka and internationally. While awareness of online shopping is growing, financial institutions must develop, to offer reliable online payment services to support the expansion of e-commerce, while creating a competitive marketplace.(“The Growth of E-commerce in Sri Lanka – Lamudi,” n.d.)
2.1.1. E-commerce market in Sri Lanka and existing e-commerce sites
Sri Lanka’s online retail industry growing fast due to good opportunity to get products at a cheaper cost, as well as lack of physical infrastructure like shops and malls, Jones Lang LaSalle, Sri Lanka said. “Not only do things look good for physical retail in the country but also for online retail,” Gagan Singh, chairperson, Jones Lang LaSalle, Sri Lanka said in a statement. “Unlike other countries where e-commerce slowly developed as a third retailing option after their high streets and malls had matured, Sri Lanka is already seeing the emergence of online retail even as its high streets and malls remain in a nascent stage.” (“Sri Lanka’s e-commerce industry poised to take off: JLL – Lanka Business Online,” n.d.).
While country has been witnessing a visible growth in per capita financial gain over the past number of years, its retail business still remains underpenetrated. The scale of the Sri Lankan retail market is estimated to vary between 25 billion to 30 billion US dollars and it is expected to deliver early double-digit growth within the short term. Sri Lanka’s e-commerce market, with external and internal travel included, is estimated at around 40 million dollars. This could grow tenfold to 400 million dollars by 2021/22. Traditional retail is estimated to grow 30 percent from 10 billion dollars to 13 billion dollars by 2021/22, according to estimates by Takas.lk and York Street Partners.(“Interview: E-commerce shakes up Sri Lanka’s retail sector, says Takas CEO – Lanka Business Online,” n.d.).
2.1.1. Existing e-commerce site in Sri Lanka
There are number of e-commerce websites in the world. Some of the major and most popular e-commerce sites are eBay, Amazon, Aliexpress, Target, Best Buy, Walmart, Snapdeal, Flipkart. These sites are open for Sri Lankans, but not only for Sri Lankans. As well as some of sellers added some restrictions to Sri Lanka. Other than that shipping cost and taxes will apply for some of items. In the last few decades e-commerce sector grows up in Sri Lanka and launched several kind of e-commerce sites. From the e-commerce sites available in the internet that relevant only for Sri Lanka, author has selected four e-commerce sites to bring out competitive edge that exists in the current market.
Figure 2 : Logo of Kapruka.com
Figure 3 : Screen Shots of Kapruka.com
Kapruka.com is one of the e-commerce type web application and Kapruka was formed in 2003 and its principals were set by a single entrepreneur. This site is not a typical e-commerce site, kapruka is mostly focus on gifts items. Kapruka is a brilliant e-commerce site that allows users to send a wide array of gifts to their loved ones no matter where are they in the world. There are over 10,000 products to choose from and over 30 types of services. The site has evolved noticeably from the inception and now offers the “Kapruka Global Shop” concept which allows customers to purchase products from international sites such as eBay, Amazon, and Best Buy without having to pay exorbitant shipping fees and allows to place the order as a single order. Kapruka was the first e-commerce site to fill the market gap of island wide gift delivery in Sri Lanka in a fast and convenient manner. (“Kapruka – Corporate Information – E Commerce Orginizations in Sri Lanka,” n.d.)
One of the major drawback of this site is lack of user behavior analytics, such as what users browsing, purchasing etc. They have large number of daily active users worldwide, but they are not using any user behavior analytics method to display suitable products for each single user, according to their behavior.
Figure 4 : Logo of Mydeal.lk
Figure 5 : Screen Shot of Mydeal.lk
Mydeal.lk is one of major e-commerce sites in Sri Lanka. It was established in 2011. MyDeal.lk recently reconstructed their website, www.mydeal.lk with a sleek, clean and uncluttered user interface, providing comprehensive deal and product related information to customers, thus facilitating well-informed online purchases. (“MyDeal.lk Celebrates 4 Years of ‘The Best Deals in Town’ – README,” n.d.).
There is one huge drawback on their site is not tracking their users behaviour. They have listed all items on their home page (random order or recent items first). Disadvantage is when two different users login to the site in same time, both users will see the same products list on their home page. This site is not suitable for the users who are not like to wasting their time for browsing.
Figure 6 : Logo of Wow.lk
Figure 7 : Home page view of wow.lk
Wow.lk started as Anything.lk, a daily deal site and today is one of the most popular e-commerce sites in the country. The site offers a wide variety of products often at a discounted rate and it’s divided into four main sections: deals, mall, lifestyle and classified. This is certainly one of the most comprehensive e-commerce sites in the sector offering over 3,000 products ranging from electronics, fashion, sport and fitness and travel. Users can make payments choosing from various methods, and the goods are delivered to them. The strategic partnership established with Dialog has allowed the site to broaden its services even wider, taking the Sri Lankan online shopping concept to new levels offering a truly comprehensive experience.
Wow.lk listed their items in category vice, such as electronics, phones / tabs and accessories, home and kitchen, computers and accessories, daily essentials etc. As other Sri Lankan e-commerce sites, wow.lk is not rearrange their content for each separate users according to their behaviour on the site. Every user login to the site in same time, they will all see the same content of the site.
Figure 8 : Logo of takas.lk
Figure 9 : Home page of takas.lk
Takas started in 2012, with a small measure of fame from the Colombo Hackathon and 6 million rupees. Takas.lk makes it easy for the consumer to buy online. Being the pioneer’s in introducing Cash on Delivery to Sri Lanka, users are able to use either cash or card to purchase. Since the inception takas.lk have grown to one of the top contenders in the e-commerce industry in the island with more than 4000+ electronics products now which their customers can choose from, and 2000+ products in areas of apparel, personal care, car care, home appliances just to name a few. Author found this e-commerce site as best and advance site compared to other e-commerce sites in Sri Lanka, because they are using user behaviour analytics on their site. If any visitor or user browse for an item or purchase an item, and after a specific point, which is around several seconds, its generate product recommendation results as a response to each activity of a user. That is, if the user visits to home page, system uses Similar Products model to select a set of product recommendation. It’s appear as a slider in each users home page.
2.2. Secondary Research
Secondary Research is mainly focused on the technologies that have been used by the author in developing the application.
Large record analytics undertaking will ingest, correlate and examine the facts resources, after which render an answer or end result based totally on an overarching query. This means human analysts need to have an in depth knowledge of the available facts and possess some feel of what solution they’re looking for.
Speed is also meaningful, as massive records evaluation expands into fields like gadget getting to know and artificial intelligence, in which analytical procedures mimic perception via finding and the usage of styles inside the amassed facts.
Similarly, security problems are magnified by using speed, extent, and a variety of huge facts, inclusive of big-scale cloud infrastructures, the range of facts resources and formats, streaming nature of data gaining, and high quantity inter-cloud migration. Conventional, legacy safety mechanisms, which might be tailored to small-scale, static (rather than streaming) statistics, are insufficient.(“What is big data ? – Definition from WhatIs.com,” n.d.)
Zettaset have advanced security solutions designed and optimized, from the floor up, to address the demands of huge information compute, storage, and cloud environments. Zettaset security solutions, along with the big data encryption suite, are enhanced for performance and scalability inside the dynamic, high-extent environments that exist in the organization and the elastic cloud.(“What is Big Data and why do organizations need it?,” n.d.)
In the end, the price and effectiveness of large facts depend on the human operators tasked with knowing the facts and formulating the proper queries to direct large data projects. Some massive data equipment meets specialized niches and permit less technical users to make diverse predictions from regular business statistics. Nevertheless, different tools are acting, together with HADOOP appliances, to help groups enforce an appropriate computer infrastructure to address huge records tasks, even as minimizing the need for hardware and distributed computing software program information.
However those equipment simplest addresses limited use instances. Many different big facts duties, along with determining the effectiveness of a new drug, can require full-size medical and computational understanding from the analytical body of workers. There is currently a shortage of information scientists and other analysts who have revelled in running with large records in disbursed, open supply surroundings.
2.2.2. Data Mining
Data mining is the method of sorting through huge facts units to become conscious of patterns and set up relationships to solve problems through facts evaluation. Data mining is also called statistics discovery and information discovery.(“What is Data Mining? – Definition from Techopedia,” 2017)
The extraction of hidden predictive records from the massive database is a powerful new generation with the high-quality potential to assist companies’ consciousness at the maximum important statistics of their information warehouses.
In data mining, affiliation guidelines are created by studying records for frequent if/then patterns, then the use of the support and confidence criteria to locate the most critical relationships within the information. Assist is how frequently the items appear inside the database, even as self-assurance is the number of instances if/then statements is correct.
Statistics mining techniques are the result of an extended process of studies and product development. Data mining is prepared for software in the commercial enterprise community because it’s miles supported by using 3 technologies that are now sufficiently mature:
- Big records collection
- Effective multiprocessor computer systems
- Data mining algorithms
The maximum normally used strategies in data mining are:
- Artificial neural networks: nonlinear predictive models that study via education and resemble biological neural networks in shape.
- Decision trees: tree-shaped systems that constitute sets of choices. These decisions generate rules for the category of a dataset. Particular decision tree strategies include classification and regression trees (cart) and chi square automated interplay detection.
- Genetic algorithms: Optimization strategies that use strategies together with the genetic mixture, mutation, and herbal selection in a design based on the standards of evolution.
- Nearest neighbor technique: a method that classifies each report in a dataset based totally on a combination of the training of the k document(s) most similar to it in a historical dataset (where okay ³ 1). Once in a while referred to as the k-nearest neighbor method.
- Rule induction: the extraction of beneficial if-then regulations from records based totally on statistical significance.
Data mining techniques are used in many research regions, inclusive of arithmetic, cybernetics, genetics and advertising. At the same time as data mining strategies are a method to force efficiencies and are expecting purchaser behavior if used efficiently, a business can set itself aside from its opposition thru using predictive analysis.
Net mining, a sort of information mining utilized in client dating management, integrates statistics accrued by traditional data mining techniques and strategies over the internet. Net mining ambitions to understand customer conduct and to evaluate how effective a specific internet site is.
Comprehensive records warehouses that combine operational statistics with the purchaser, supplier and market information have resulted in an explosion of facts. The competition calls for well-timed and sophisticated evaluation on an integrated view of the facts. But, there’s a growing hole between the more effective garage and retrieval systems and the customers’ ability to efficiently analyze and act on the records they comprise. Each relational and OLAP technologies have outstanding abilities for navigating massive facts warehouses, but brute force navigation of statistics isn’t sufficient. A new technological jump is needed to shape and prioritize data for specific stop-person troubles. The facts mining gear can make this leap. Quantifiable enterprise benefits were proven through the integration of statistics mining with modern statistics structures, and new merchandise is at the horizon so one can bring this integration to a fair wider audience of users.
Many data mining equipment currently function out of doors of the warehouse, requiring greater steps for extracting, uploading, and studying the facts. Moreover, when new insights require operational implementation, integration with the warehouse simplifies the utility of consequences from data mining. The resulting analytic facts warehouse can be carried out to enhance business strategies during the business enterprise, in regions including promotional campaign control, fraud detection, new product rollout, and so forth.
The right place to begin is a data warehouse containing a combination of internal information monitoring all consumer contact coupled with external marketplace information about competitor pastime. Background records on potential clients also give the first-rate basis for prospecting. This warehouse can be implemented in a variety of relational database structures: Sybase, Oracle, red brick, and so on, and have to be optimized for flexible and speedy statistics get admission to.(“An Introduction to Data Mining,” n.d.)
An OAP server allows an extra sophisticated stop-user commercial enterprise version to be applied when navigating the information warehouse. The multidimensional structures allow the person to analyse the records as they need to view their business – summarizing via product line, vicinity, and different key perspectives in their business. The records mining server ought to be incorporated with the information warehouse and the OLAP server to embed ROI-centred enterprise evaluation at once into this infrastructure. An advanced, manner-centric metadata template defines the facts mining goals for specific business issues like campaign management, prospecting, and advertising optimization. Integration with the statistics warehouse permits operational selections to be directly implemented and tracked. Because the warehouse grows with new choices and outcomes, the employer can always mine the exceptional practices and follow them to destiny choices.
This layout represents a fundamental shift from conventional decision assist structures. as opposed to genuinely delivering facts to the end consumer via question and reporting software program, the superior evaluation server applies users’ enterprise models immediately to the warehouse and returns a proactive analysis of the most relevant records. Those effects enhance the metadata inside the OLAP server through presenting a dynamic meta data layer that represents a distilled view of the information. Reporting, visualization, and different analysis equipment can then be applied to devise future moves and verify the effect of these plans.
2.2.3. Machine Learning
PHP-ML is a new approach to Machine Learning in PHP. PHP-ML library includes Algorithms, Cross Validation, Pre-processing, Feature Extraction and much more in one library. PHP-ML requires PHP 7.0 or higher version and available under an MIT license. Currently, this library is the technique of growing, but it could install with the composer. Types of using PHP-ML, machine learning library are affiliation rule learning as well as classification. In Classification train and predict techniques are identical. Algorithm used in classification method is classifier.
Figure 10 : One of the simple example of classification
The classifiers supported in PHP-ML are SVC (Support Vector Classification), K-Nearest Neighbors and PHP-ML also uses a feature Naive Bayes .Bayes theorem is the base of Naive Bayes. Parameters are not included in Constructor in this feature.
o Decision Tree (CART)
o Ensemble Algorithms
o Bagging (Bootstrap Aggregating)
o Random Forest
o Ada Boost
Regression is the important feature in PHP-ML which is continuous value .Multiple linear regression is also the option given by PHP-ML.
- Least square value
Linearity is also a type which includes:
o Decision Stump
o Logistic Regression
Grouping of identical objects is named as clustering in PHP-ML .It supports
Fuzzy C- Means
o Confusion Matrix
o Classification Report
Neural network supports Multilayer perceptron classifier .Feature extraction of PHP-ML supports Token Count Vectorizer ,Tf-idf Transformer .Datasets of PHP-ML includes Arrays ,Files,CSV. Three predefined datasets are available to be used in the library which are Iris, Wine and Glass.
- Database Selection
A database is a prepared series of records. Its miles the gathering of schemas, tables, queries, reviews, views, and other items. Since their advent in the 1980s, relational database control systems (RDBMS) have ended up the standard database kind for a spread of industries. As their name implies, those structures are based totally on the relational version that organizes records into agencies of tables referred to as relations
Famous DBMS include MySQL, Microsoft SQL SERVER, Oracle, Sybase, SAP HANA, SQLite and IBMDB2. But the author needed to select one database system to store the information about users, vendors, destinations, accommodation, transportation and other important facts while growing the e-commerce website online.
22.214.171.124. Comparison of databases
In this section we do a comparison between three famous RDBMS,
The first “actual” launch of the Oracle RDBMS changed into Oracle 2. This device supported simplest basic SQL. Capabilities and it was written in an assembly language. The latter year, and for the later 10 years or so, Oracle Company released updates to its flagship database. Probably one of the motives that the Oracle RDBMS has controlled to stay at the pinnacle of powerful RDBMS is linked to its product updates which can be intently tied to changes in the marketplace. Database buzzwords which include “scalable”, “programmable”, “allotted”, and “portable” also are tied to Oracle launch. As an instance, in 1985 and for a client-server version became brought in anticipation of a growing attractiveness of community communication. Because the internet paved the way for the virtual era, the Oracle RDBMS changed into up to date to include a local JVM.
Oracle Database 12c is the most recent launch of the RDBMS, and it includes the following functions:
- New records reduction to enhance safety of sensitive statistics
- Introduction of Oracle superior analytics platform
- New database dealing with for archiving flash data archive (FDA)
- Guide for integrating with operating machine processor groups
- Help for information pump for database consolidation
- Superior network compression to beautify performance
While, in 2000, Microsoft released SQL server 2000. The release changed into a tremendous milestone for the company because it marked the first launch of the product where the unique Sybase code changed into completely new code. Within the equal vein as Oracle organization, Microsoft has attempted to decorate square server to keep up with converting generation. SQL server 2005 is an example. The XML received the stamp of approval from w3c and commenced gaining ground in the past due Nineteen Nineties. One of the fundamental new capabilities of square server 2005 was support for XML statistics. Other great features of the flagship product consist of the creation of square server always on (facts management technology to lower user downtime), assist for dependent and semi-dependent records, stronger compression, and numerous add-ons to assist different merchandise on the market. SQL server 2012 became proclaimed because the remaining launch to include a local guide for one. A SQL server 2012 essentials course can provide greater statistics about this platform and how to use it.
SQL server 2014 is the latest release of square server and consists of the subsequent functions:
- Advent of in-memory OLTP, an embedded characteristic that permits sophisticated database management to beautify performance
- New solutions to handle catastrophe restoration
- Updated version of square server data gear for business intelligence
Oracle and SQL server are considered gear that wants users with large business enterprise systems, whilst MySQL is considered a tool that appeals most usually to individuals interested in coping with databases associated with their web sites. As with Oracle and SQL server, MySQL has launched updates to its software program just about every year. The original model became evolved within the mid- Nineteen Nineties. The maximum exceptional changes to MySQL turned into in 2010, the time of the remaining acquisition in 2010. The enhancements to this launch (GA release 5.5) blanketed semi-synchronous replication, custom partitioning, progressed assist for SMP and updates to the INNODB I/O subsystem.
Microsoft SQL server makes use of C++.At the same time as in MySQL and Oracle implementation languages are c and C++.
Secondly, Microsoft SQL server and Oracle are industrial while MySQL is open source. SQLite is for embedded structures, facts analytics and so on. SQLite is a software library that implements a self-contained, server less, zero-configuration, transactional square database engine. SQLite is the most extensively deployed database engine in the world. SQLite isn’t always as lots of a SQL server. Its miles absolutely just a database tool. the databases may be embedded in applications, or used domestically without an extra “server” resources as its accessed greater like a textual content report would be accessed than a server. Nearer merchandise would be Microsoft SQL CE. SQLite is a small version of MySQL.
This contrast indicates just how near the databases are in three key areas. Considering your precise scenario is probably greater applicable for figuring out which one to enforce than figuring out which one is first-rate.(“Microsoft SQL Server vs. MySQL vs. Oracle Comparison,” 2017)
126.96.36.199. Evaluation on database
Among Microsoft SQL server Oracle and MYSQL, author determined to apply the MySQL database in developing the solution. The main cause for selecting the MySQL database is due to the fact it’s far open source RDBMS. When thinking about the recognition it’s far within the second rank. Maximum of the capabilities are identical in all 3 databases and there are no boundaries in the MySQL database due to the reality that it’s far open source. Some different blessings of using MySQL database are ease of use, safety, efficiency, scalability, flexibility, high availability and so on. This is the maximum strong database tool that’s sincerely free to download and set up.(“Oracle vs. MySQL vs. SQL Server: A Comparison of Popular RDBMS,” 2013)
Figure 11 : Logo of Cxense
Cxense offers a unified and collaborative solution that allows its customers and partners to effectively interact with their online audience. This process is done with the context and analysis of this context (the combination of many variables such as time and place, device, role, preferences, content, intention, event, activity and so on) in which content is consumed and produced, the context in which entities and concepts are used within the content and perhaps most importantly, the context of the users themselves.(“Cxense – Gigya Documentation – Developers Guide,” n.d.) Cxense is a company that is provide a platform for advertising, data management, browsing, analytics and content recommendations. This company headquartered in Oslo, Norway, with branches in Boston, Buenos Aires, London, Madrid, Melbourne, Miami, New York, Rio de Janeiro, Singapore, Stockholm, and Tokyo.
We can purchased cxense platform and integrate it with our website. They have provided an analytic script that we need to deploy all of the pages that need to measure interactions in our website. Then we will able to personalize the content of the site for every single user.
Available at https://www.cxense.com/
(“cxense – Wikipedia,” n.d.)
Figure 12 : Logo of Gigya
Gigya enriches Cxense Analytics with personal, professional and social user information such as birth days, name, address, skills, education, favorite movies, hobbies, patents, Facebook likes etc.
Figure 13: Logo of Evergage
Figure 14: Logo of Google Analytics
Google analytics (GA) is a free service for web analytics that provide by google. That provides statistics and basic analytical tools for SEO and marketing purposes. This service can use anyone with google account. GA features include Data visualization tools including a dashboard, scorecards and motion charts, which display changes in data over time, Segmentation for analysis of subsets, such as conversions, Custom reports, Email-based sharing and communication, Integration with other Google products, such as AdWords, Public Data Explorer and Website Optimizer. GA is geared toward small and medium-sized retail websites. The service has limitations that make it less suited to more complex websites and larger enterprises.
Webtrends Infinity Analytics
Figure 15 : Logo of Webtrends
The latest generation of Webtrends Analytics is Infinity Analytics. Built on the data-streaming Webtrends Infinity big data platform, Infinity Analytics delivers all of the insight that need to manage and measure digital business in a highly interactive with user-directed manner. While typical digital analytics solutions are forced to balance trade-offs between scale, speed, accuracy, and interactivity, Infinity Analytics delivers all these benefits without trade-offs. To improve accuracy, some other analytics tools have offered speed of results and response time of data. Such as Webtrends Analytics 10 products that did effectively manage both scale and accuracy, but that were not able to provide everything in real-time. And when data was delivered, it was not co-operative, but typically a more static view, due to its immense processing required to deliver both scale and accuracy. Infinity Analytics is able to slove these problems, delivering interactive, highly scalable and accurate results quickly.
With webtrends Infinity, data is streaming into a highly secure globally distributed data collection network continually. After data is collected, it is sessionized, evaluated, and augmented to identify if someone is a new user, or previously known user. The sessionized data rapidly streams into a Hadoop data lake where it is instantly available for real-time query. Results can be streamed to the Infinity Analytics application, or out for integration to other marketing systems to drive personalization programs. A true big data architecture, Infinity Analytics leverages the latest highly scalable big data and streaming technologies such as Hadoop, Spark and Kafka. These leading technologies offer unprecedented speed and scale, enabling a significant degree of interactivity and on-the-fly analysis.
(“Infinity Analytics | Product Help,” n.d.)
Figure 16 : The process of Webtrends Infinity
Chapter 4: Requirements analysis
This chapter encompasses about the analysis of requirement that author identified of dynamically content rearranging e-commerce web site. Requirement analysis is a highly valued section to start of the development lifecycle, which helps to identify what needs to be done and get a rough idea of the domain of the project. For this analyse we use information that discovered in the literature review chapter (chapter 2) and current system analysis which in related projects (chapter 3) to identify what we have to do to finish this project. Firstly need to identify stakeholders of the system before applying standard requirements analysis techniques. After that it is possible to generate use cases and user stories. Finally the function and non-functional requirements can be stated. Requirement analysis phase plays a major role on developing, because it shows that the development is based on exclusively around a user or a client requirements rather than based on what developer wants to do. We will be use the requirement that we specify in this section to create test cases in testing chapter.
4.1. Stakeholder Identification
This is the first phase of requirement analysis technique that we use. To get a rough idea of the system we have to identify all users or stackholders that using the system or interact with the system. When consider about our system, it is an e-commerce website. The main stakeholder will be the user who will use the e-commerce site, will refer them as the end user (customer). There are two types of end users, they are visitor and registered user. There is an administrator type stakeholder who is solely interact with the system would be database management. Anyway major focus is on end user, because end users are the greatest participation in this system.
3.2. Use Cases / User Story
Use cases are subset of UML diagrams. A use case diagram is a description, which include a collection of interactions between a system and one or many actors. (“Use cases vs user stories in Agile development | Boost Blog,” n.d.). User stories are not use cases. By themselves, user stories don’t offer the main points of the details the team needs to do their work. User story include a short description of something that use may do after they come to the website or web application. It focused on the value or result they get from doing this thing.
In UML diagram stakeholder identified as an actor and it shows as a stickman. Actions and methods visually depict by a circle and each line shows which actor performs which action. In some of cases action can extend another or include another, extend is used when a use case adds additional steps to other first class use case and include is used when there is any identified common behaviour in two use cases. The included use case is not able to keep stand alone and the included one must need to complete original use case.
Figure 17 : User case for the e-commerce web application
Below are the use case narratives for each action in the use case diagram:
Table 1: Use case scenario for View Products