×
Community Blog GreenBytes: Reducing Food Waste in Restaurants with AI

GreenBytes: Reducing Food Waste in Restaurants with AI

This project is from the team of GreenBytes, which was awarded third place (shared) in the Alibaba Global AI Innovation Challenge 2020.

1

This project is from the team of GreenBytes, which was awarded third place (shared) in the Alibaba Global AI Innovation Challenge 2020.

Project Introduction

GreenBytes is here to combat food waste in restaurants. We have engineered a cloud-based solution for restaurants to save money and reduce waste by letting them know how much food they should be ordering. We have developed a progressive web application that breaks down menus, tracks inventory, and predicts future food consumption using machine learning algorithms. More specifically we use a recurrent neural network that inputs past sales to predict future sales. Once we predict future sales, we suggest how much of each raw ingredient the restaurant should order in the upcoming days. If the restaurant agrees with our suggestion, they can approve the order and we will automatically send out the order to all of their distributors.

Target Problems

Understanding trends in restaurant sales by looking at past sales and various weather patterns allows GreenBytes to reduce almost 2 tons of food waste in one month in just one restaurant. This is important because food waste produces a huge amount of greenhouse gases, impacts our land, our water, our biodiversity, and our livelihood in general.

Worldwide, approximately 3.3 billion tons of CO2 are released into the atmosphere due to food waste each year. Food waste emits more GHG emissions than any country other than China and the U.S.

GreenBytes prevents food waste due to expired food, too much food prep, and table leftovers by predicting how much of each menu item will be ordered in the future, making ordering suggestions, and tracking inventory.

Our Solution

We have developed a progressive web application that breaks down menus, tracks inventory, and predicts future food consumption using machine learning algorithms. More specifically we use a recurrent neural network that inputs past sales to predict future sales. Once we predict future sales, we suggest how much of each raw ingredient the restaurant should order in the upcoming days. If the restaurant agrees with our suggestion, they can approve the order and we will automatically send out the order to all of their distributors.

To predict future food sales in restaurants we consider past sales, weather data, and COVID-19 statistics.

Alibaba Cloud Products Used

Technology Highlights

Two approaches were used to analyze the data for this project. One approach used PAIs' built-in time series analysis and another used deep learning.

PAI was used to apply the deep learning algorithm. The python model and training set were uploaded to OSS. The TensorFlow component was used and the pat to the python code files and data source directory were set.

The remainder of the project was completed outside of PAI using python to complete the post-processing. The post-processing included taking the results from the ML models in PAI and performing calculations with the menu breakdown to determine how much of each ingredient was necessary for each day.

About the Team

Our three-person team has background in computer science, data analytics, engineering, sustainability, and food retail. The GreenBytes team is made up of Renata, Jillian, and Ilona. Along with our strong technical backgrounds we have over a decade's worth of experience in the foodservice industry and learned about business development through various start-up accelerators. Our hands-on experience in restaurants gives us a tangible understanding of the food waste problem and our technical know-how gives us the ability to help solve the problem. Apart from our experience, we are resourceful. We thrive in chaos and embrace new challenges with curiosity. Our passion for the planet is our why and our problem-solving abilities is our how.

0 0 0
Share on

Alibaba Cloud Project Hub

57 posts | 10 followers

You may also like

Comments

Alibaba Cloud Project Hub

57 posts | 10 followers

Related Products