To date, anaplatform has proven itself with a state-of-the-art e-commerce platform which is a software solution suitable for different customer needs with many analytics capabilities. Our customers, among them iMyShop http://imyshop.net/ , the shopping giant of the Middle East. If you want to develop and improve the e-commerce programs you have, we can solve it in terms of integration, or we write from scratch to suit your needs. We also provide support by analyzing your architectural requirements. With the introduction of e-commerce systems integrated into trade, competitiveness has increased, and it has made significant contributions to the growth of companies and increase in trade volumes regardless of time and place

While a decrease was observed in general trade during the beginning of the Covid-19 epidemic, it was clearly seen in the statistics that there was no such decrease in e-commerce and that e-commerce continued its increasing trend. Contrary to classical commerce, e-commerce continued to rise with the effect of the pandemic and the campaigns of e-commerce sites.

Customized E-marketplace

With our platform, which will appeal to a wide audience, especially food companies, the expenses of the store owners will be minimized and the maximum number of potential customers will be reached. In this way, a marketplace was created that will enable the companies in different regions, which come together through e-commerce, to grow in the fastest way and to enable users to access the products they need at the most affordable prices. Our project product is a software product that can represent our country as a Turkish e-marketplace in the local and international arena, with a wide variety of products and many companies. The contribution of e-marketplaces, which is a commercial area that can survive under all conditions, to the trade volume is quite large.

As a result, Electronic Commerce adds dynamism to the sectors, thanks to the new business lines it has created, new employment and expertise areas, and new markets it has reached. Although Turkey does not yet have a significant share in the global e-commerce pie, it has a significant potential in e-commerce, which is an important part of a digital transformation supported economy, as a country that adopts and evaluates the opportunities offered by the internet more and more.

Our e-commerce platform is an e-commerce platform that will be offered to the user on web and mobile platforms. Our platform is predominantly developed using artificial intelligence, natural language processing (NLP) and cloud technologies.

The latest possibilities of cloud technology have been used, thus providing access to infrastructure that can reach more users quickly with less expense. In addition, in this way, damage to the environment will be avoided by installing more servers than necessary, and energy savings will be achieved. Within the scope of our application, there are also Marketplace panels and admin panels to allow store owners to sell their products.

Semantic Search

With the help of Natural language processing Semantic Search is used in the product search within the platform. In this way, even if the user searches with a long sentence on the search screen, the sentence is processed, the words are cleared of attachments, etc., and the database is searched by getting to the root of each meaningful word.

In this way, convenience is provided not only to the users who buy the products, but also to the people who make sales and manage the content of the site. While the features mentioned so far will ensure that our platform has a place in the market, our more advanced recommendation systems that we will add to our platform with artificial intelligence technologies will enable us to stay ahead of our competitors.

Next Generation Recommendation Engine

We have developed a state-oft-the-art recommendation engine with a moderated approach within the platform. Thanks to our recommendation system using machine learning technologies, users can see the products that most interest them and find the product they are looking for without wasting time. Due to the use of artificial intelligence technologies for different purposes, the platform will be in a continuous development with the interaction of different artificial intelligence sub-branches with each other and with web services.

A recommendation system has been developed with the big data obtained by collecting the data of the users using artificial intelligence technologies. With this, it is possible to bring the most sought-after, the most popular options to the user, which they can choose according to the user's orientation. While this saves time, it also replaces advertising support. For example, a customer has the opportunity to see the prices of the product he examines at various vendors, while he is aware of the existence of similar products. This increases the display rate of the products of the sellers and thus the sales rate.

While the suppliers using our system can track the preferences or behaviors of a limited number of customers who visit their own website and shop, our e-marketplace application, which hosts transactions between millions of consumers and tens of thousands of sellers every day, provides users with a unique experience that their competitors or even sellers who sell on the platform do not have. provides access to data. What products and sellers a consumer is interested in, which brands he spends more time on, what kind of product he buys, how much he spends, his approximate income, etc. has a lot of data. This provides a competitive advantage for the company with the data.

In addition, according to the payment choice of the users, the most suitable pricing among the payment system providers will be dynamically selected, thereby increasing the income of the business owners.

While developing the platform, AWS Services were mainly used on the cloud computing side. Among these services, AWS Cognito was used for secure authentication and authorization (user systems), Amazon S3 for object storage, and Amazon OpenSearch Service for fast word-based searches and word completion to search for required products. In addition, by taking advantage of the serverless architecture on Amazon, we will avoid hardware maintenance costs in our project, and for pricing, a pay-per-use method has been preferred in many of our API services.

Ecommerce Analysis

RFM Analysis, Basket Market Analysis and various Regression Analysis methods were used to provide automatic customer segmentation and customer analysis. A recommendation engine with a supervised approach (Supervised Learning) has been developed in the recommendation system. Real Time Sentiment Analysis provided by auto collection of review from customer comments by Web Scraping methods. For turkish senthesis, Zemberek library is used for natural language processing in the background in the search section. Zemberek is an open source Turkish Natural language processing library.

Our APIs are also available on Amazon web services in order to ensure communication with the front side of our application and to minimize costs. Not only the REST API but also GraphQL is used. In this way, the information needed on the front face can be accessed with an optimized data load, thus reducing our application performance and costs. Since the project product has a complex structure,

Serverless

architecture structure has been used and the advantages of both SQL and NOSQL database systems have been utilized. In addition, AWS SNS and AWS SES services were used to inform the user about the opportunities about the products they were interested in, and notifications and personalized e-mails were sent.