Reducing waste in food production
Context & Objectives
A common goal for nearly every food production company is to minimize waste. Whether it is to build a sustainable business with minimal environmental impact or to reduce overheads, there is no doubt that companies stand to gain a lot by reducing their waste. This was no different for a large food production company based in France.
To do this, they recognized that they needed to take a data-driven approach. But where do you even start? If you're in a similar situation, you may have large volumes of data from every stage of your supply chain process located in many different databases and warehouses. At a certain point, the volume and complexity of the data cause this investigation to be more effort than it's worth, right? Well, that's where data science and the Agilytic team come in.
The main objectives are:
Consolidate the different data sources and gain an understanding of the entire data ecosystem.
Provide insights into which factors are responsible for the majority of the waste in the production process.
Develop a roadmap with the actions that must be taken to better deal with waste.
Approach
Before any analyses could take place, around a dozen different data sources needed to be aggregated and cleaned. Data ecosystems of this size usually come with many challenges. One of the biggest challenges faced here is the improper classification of waste at various stages of the supply chain. This caused duplications and overestimated the true quantities of waste produced.
Understanding the data ecosystem through the activities in this first step alone is enough to shine a light on possible shortfalls in the supply chain process, and this company can take action immediately to rectify them.
Next, we use machine learning to identify the factors responsible for most of the waste in the production process. Through this analysis, we gain insight into the impact of several factors on the quantity of waste produced, such as
The brand and type of product produced.
The lifespan of each type of product.
The time required to sell a product.
Many factors used in the analysis needed to be carefully constructed from the available information. Often, the best indicators of our outcome (in this case, waste) are not visible or available and must be engineered.
Results
Through the analyses we conducted, this food production company was able to:
Get a better understanding of their data ecosystem to build a more reliable foundation for future analyses and forecasting.
More accurately quantify their waste so that action can be taken in the right place and at the right time.
Identify the factors that contribute the most towards their waste so that they can take action to address them.
Determine which products and processes they need to prioritize and improve to minimize waste.
—
Written for Agilytic by Joleen Bothma.
