Data Science for Materials - a Collection of Open Resources for Education
Author: Dane Morgan
Date Uploaded: 2025-04-03
Content Type(s): This is a full course with lecture slides, labs, HW, and project guidelines
Content Length: Materials for one semester
Content Audience: Accessible to any level, UG or Grad
Content Topics: machine learning, databases, featurization, neural networks, convolutional neural networks, and worked examples
This course provides students with an introduction to applications of data science in materials science and engineering, including many closely related concepts from chemometrics. Topics include understanding the creation and use of modern data resources, data-centric approaches to materials, and the integration of machine leaning across the materials landscape. Some specific areas of focus are the use of databases in materials discovery and design, the application of machine learning to materials property prediction, analysis of characterization data of images, machine learning interatomic potentials. We go from basic introduction to state-of-the-art, e.g., basic data prediction workflows to Transformers. The course is python based and application oriented, with a goal of learning to effectively use tools at an advanced level. A large part of the course is centered on hands on computational laboratories and a final project. The course is targeted for upper-level undergraduate and beginning graduate students, but requires no particular background. Prior familiarity with programming, machine learning, computer modeling, etc. is not required but the more relevant background one has the easier the course will be.
It is very hard to use a course without some information about how the materials were used - feel free to reach out to me to discuss! Dane Morgan <ddmorgan@wisc.edu)