Machine Learning (ML) encompasses a large and growing number of methods and concepts that are increasingly relevant across disciplines and industries. To make these concepts tangible in teaching, we have developed eight diverse ML use cases at different difficulty levels — beginner, intermediate, and advanced. Each use case is based on authentic industrial or scientific data and addresses real-world problems such as classification (e.g., Trumpf, Blueberry), regression (e.g., Fiber Content), or time-series forecasting (e.g., Sales Data). These challenges enable students to experience realistic ML workflows — from data exploration to model evaluation — while understanding both potentials and limitations of ML approaches.
The use cases serve as an entry point for hands-on learning: they allow students with varying backgrounds in ML, statistics, or programming to apply theoretical knowledge to practical problems. By engaging with real datasets rather than artificial examples, learners strengthen their data literacy, gain insight into how ML can support decision-making processes, and learn to critically reflect on issues such as bias or uncertainty in data-driven models. Working collaboratively on these tasks also promotes interdisciplinary cooperation skills. The topics cover domains ranging from engineering to business applications, encouraging teamwork among students with different academic perspectives. This approach supports competence development not only in technical aspects but also in communication, problem solving, and critical thinking.
The table below provides an overview of all available use cases, including their task type, difficulty level, and a short description. Each use case has its own webpage with detailed information about the dataset.
All materials are designed for flexible integration into university courses — suitable for both distance learning and on-site teaching. To support educators, we provide model solutions for all challenges. Please contact us via email (lara.kuhlmann@tu-dortmund.de) if you would like access to the datasets or further information about integrating these use cases into your teaching activities.
| Challenge | Task | Difficulty | Short Description |
| Trumpf | Classification | Intermediate | Prediction of the successful removal of sheet metal parts. |
| Sales Data | Time Series Forecasting | Beginner | Monthly sales forecast for five products. |
| Hand-drawn unit operations | Classification | Advanced | Classification of hand-drawn process engineering symbols. |
| Molecular feature prediction | Regression | Advanced | Prediction of boiling point based on chemical properties of a molecule. |
| Blueberry | Classification | Intermediate | Classification of blueberry production into three levels based on cultivation data. |
| Fiber content | Regression | Intermediate | Prediction of Arabinoxylan content from cultivation and weather data. |
| Public procurement | Classification | Beginner | Classification regarding savings. |
| Public procurement | Regression | Beginner | Prediction of a purchase price. |