Course Description: Informatics is the use of computational tools (e.g. AI, machine learning) to access and make sense of data. Materials Informatics has the potential to accelerate scientific inquiry and support novel insights. However, informatics tools are typically marketed to the biological sciences, financial sector, and “Big Tech”---it is not immediately clear how to use “cat-detector technology” to support serious materials science! This hands-on workshop will help material scientists use programming-based informatics tools to create reproducible data workflows, automate tedious data extraction, visualize data for exploration, and apply AI/ML models for predictive capabilities.
Topics Covered: (Coding) Python & Jupyter notebooks; (Data extraction) Tabula, WebPlotDigitizer; (Data handling) Pandas / Grama; (Visualization) Plotnine / Plotly; (Modeling) Over/under-fitting, cross-validation, domain of applicability, active learning.