Data science and data-driven methods are an essential component of modern engineering and affect all stages of the product lifecycle, from data-supported product design to production, quality assurance, and intelligent or autonomous technical systems. This module introduces fundamental concepts, methods, and workflows of data science and machine learning with a strong focus on engineering applications.

Students gain an understanding of engineering-specific data characteristics and learn how to design and implement basic data analysis and machine learning pipelines. One component of the course is to work on two data science and machine learning challenges using engineering data sets. Students apply state-of-the-art Python libraries to explore data, build models, and evaluate their performance.

The topics of this course include:

• Engineering data sources, data types, data quality, and data representation
• Data science workflows, exploratory data analysis, and visualization
• Statistical modeling and machine learning methods (regression, classification, clustering)
• Model training, evaluation, validation, and generalization
• Overfitting, model selection, and hyperparameter optimization
• Machine learning pipelines and practical implementation aspects
• Fundamentals of deep learning and neural networks for engineering applications

This module provides foundational competencies for subsequent courses in data-driven engineering, machine learning, and artificial intelligence.

Learning Objectives

After successful completion of the module, students are able to:

• Explain the role of data science, machine learning, and deep learning in engineering and their complementarity to simulation- and optimization-based approaches.
• Analyze engineering data sets with respect to data quality, structure, and suitability for data-driven, learning-based methods.
• Apply exploratory data analysis, statistical modeling, machine learning, and basic deep learning methods to engineering problems.
• Design and implement data analysis, machine learning, and deep learning pipelines using state-of-the-art Python libraries.
• Evaluate the performance, generalization capability, and limitations of machine learning and deep learning models using appropriate validation techniques.

Organizational Information

See further details on course organization at: lehre.imi.kit.edu and in ILIAS

Workload

120 hours

The exercises take place on Wednesdays on specific days from 9:45 a.m. to 1:00 p.m. For further information, see ILIAS.