Abstract

It can be seen that the brewing industry generates significant by-products such as spent grains, hops, and yeast, which are often underutilized or disposed of inefficiently despite their potential value. To align with the Circular Economy (CE) principles, which require manufacturers to keep resources in use for as long as possible, breweries must transform by-products into value-added resources, thereby contributing to sustainable development by reducing environmental impacts and resource depletion. While brewing factories recognize the importance of waste management and by-product reuse, there is a lack of comprehensive studies on predicting by-product quantities. This gap hinders efforts to optimize resource usage and reduce waste disposal.

 To address this challenge, this paper aims to optimize waste management and enhance the reuse of brewing by-products, contributing to the sustainability goals of the CE. Machine learning (ML) models are applied to predict by-product quantities. The models’ performance is then evaluated using regression metrics, with Gradient Boosting emerging as the most accurate. At the same time, the dashboard is employed to visualize the predicted and actual by-product data. The study uses real-world data, covering multiple production variables over five years, ensuring industrial relevance and empirical validity. The research assesses prediction accuracy and interprets how predictive insights can be integrated into daily brewery operations through dashboard analytics, enabling data-driven decision-making.

The resulting insights enable breweries to forecast waste volumes, adjust production, optimize operations and reduce inefficiencies. Furthermore, this approach provides a scalable framework for other food and beverage manufacturers seeking to apply predictive analytics for sustainability purposes. The study also highlights the potential for data-driven models to improve resource management in the brewing industry, fostering sustainable practices and advancing the goals of the CE.