Building a Forecasting Model to Predict Sales of Fresh Food
In that two-month AI challenge, a global team of 50 changemakers joined to help a foodtech startup further build its AI solution to prevent food waste.
Today, the world wastes or loses around a third of the food it produces, while almost 690 million people are hungry. If we keep throwing away 1/3 of the produced food, we’ll need 2 planets to feed 10 billion people in 2050. Reusing and recycling food waste is only a part of the solution. It’s the prevention that we need most to build a sustainable food future. According to Project Drawdown, solving food waste is the best solution for climate change.
Sustainable Development Goal 12: Ensure sustainable consumption and production patterns, contains a wide range of targets, one of which is closely related to halving global food waste at the retail and consumer levels. To feed the world sustainably, global initiatives, activities, and projects are held on food losses and waste reduction including this project.
The project outcomes
The team has built a cutting-edge ML model that accurately predicts sales of fresh food in retail, leading to better ordering and less waste.
We’ve also developed a powerful scraper that collects promotion data from retailers in Belgium, covering both fresh and non-fresh products. And with our sophisticated ontology, we can unify data from various sources through semantic matching while maintaining records of original product and promotion names. Plus, our model links web-based promotion data to sales data, giving you a comprehensive view of your operations.