Smart-List helps you save money on your shopping and prevent food waste

Smart-List is a smart shopping list that uses machine learning and deep learning methods, aiming to create shopping-conscious societies by using innovative methods in order to leave a sustainable world to future generations and reduce excess consumption.

It aims to make the user adopt the “take as much as necessary” policy.

Project logo

In addition, the Smart-List project, supported by TÜBİTAK and ESTU BAP, competed as a finalist in the field of Environment and Energy Technologies at TEKNOFEST 2021.

In addition to the conveniences provided by e-commerce sites today, people are pushing for unnecessary expenditures and unless this habitual behavior is left, food shortages and economic problems await future generations in the long run.

The Food and Agriculture Organization of the United Nations (FAO) also predicts that 300 million people will struggle with hunger in 2050. In addition, studies show that most of the food waste is done in our homes. One out of every 4 breads bought into homes is thrown away every day. Today, it is of great importance to gain conscious consumption habits in preventing food waste.

User friendly mobile application

The Smart-List application is a user-friendly mobile application that provides a personalized possible shopping list for the next purchase based on the user’s purchases in the previous months.

This application provides users with specific product and quantity information and supports purchasing in line with their needs, while changing the shopping habits of users, aiming for more conscious consumption, less food wastage and savings. This smart list provides a personal list by predicting the products according to the needs of the users after being trained with the machine learning model with real user data.

Users can track what and how much they will buy, thanks to this list, which is updated in real time as they shop, and they can also control their budget as they can record their expenditures on the application. In this way, it is aimed to reduce the consumption of products beyond the needs of the users and to enable them to gain awareness of conscious consumption.

The machine learning base of Smart-List was developed on Colab with Python programming language. The data used in the project were obtained from Marketyo company, which provides online shopping service for local markets, and these data obtained from sales in Ankara covered a 6-month period.

In line with the purpose of the project, the first 5 months of data of the user was trained by using the deep learning model Long-Short Term Memory (LSTM), which is an effective method that is also used to understand long-term dependencies. As a result, a total of 50 new estimated product outputs were obtained from this model, consisting of 2 features as product and quantity information.

The accuracy of the prediction outputs was measured by comparing this output with the user’s data allocated as test data. In tests conducted on 11 different users, an average of 75% success was achieved.

Since Smart-List provides convenience to the user and will bring customer continuity, it is thought that integrating it into online shopping sites will be very beneficial for both the user and the company.

Team Ecem Şen, Çağan Bıçakçı and Dr. It consists of 3 people, including Burcu YILMAZEL.

This project, which was started as a thesis study in 2020, received support from TUBITAK and ESTU BAP. It also competed as a finalist in TEKNOFEST 2021 in the field of Environment and Energy Technologies.

Related Posts
Read More

Unique Turkish technology in natural colorants

PICTA develops and produces special products for buyers such as herbal extracts (extracts) with high added value from agricultural materials and natural colorants. The company, which makes Turkey the world's top quality red beet producer with Islah R&D, is also involved in the production process.
We use cookies on our website to give you the most relevant experience by remembering your preferences and repeat visits. By clicking “I understand”, you consent to the use of ALL the cookies.