Client'schallenge
This specific client delivers freshly product to their customers. One of the challenges of the company is to understand food demand in stores and restaurants. Twice a week, restaurant managers are required to order products for next day delivery. This is a tricky task for any managers to simply anticipate the need of their clients.
The goal is to help restaurant managers in forecasting the demand, and to achieve a good balance between food waste and shortage.
The Solution
To better predict required stocks in stores and restaurants, a Machine Learning model (ML) was developed, to learn from sales tickets data. In addition, a deep learning neural network was developed to predict the quantity, per product, that individual restaurants should purchase for the coming days.
The overall solution was fully integrated and automated within a (Microsoft Azure) cloud environment.
Created impact
Thanks to the fully automated solution, a more accurate forecasting on food demand within this client’s restaurants environment has been obtained. Additionally, the result is more than just a tool to assist restaurant managers; this solution also allows for food waste reduction and increase in revenue.
