AI & WASTE REDUCTION

Those handy dishwasher capsules we all use come packed together in a plastic bag, not individually wrapped. But here's the catch: making these plastic bags isn't always straightforward. Production issues often lead to quality problems, meaning tons of plastic end up wasted.

This project focused on identifying the factors behind these packaging problems. First, the relevant parameters that might affect the plastic bags' quality were identified with the help of experts. Then, raw data was extracted from the packaging machine, transformed into useful features to reflect the relevant parameters. To find the best results for the problem at hand, various models were created, including logistic regression, neural networks, and gradient-boosted trees.

The interpretation of these models identified key issues, allowing for a relatively simple technical solution. This solution has since been implemented, leading to savings of €20,000 and a reduction of 5 tons of CO₂ emissions each year.

Python | JMP | AI & ML | Data Analysis