Patterns Matter is my personal toy database mapping "Raw data - Featurized data - Results" across the interface of Materials Science & AI. It started as an ML project archive and has grown into a home for my broader computational work — DFT, molecular dynamics, ML surrogates, and coupled DFT-ML-FEM pipelines. Data-driven insight, in Materials or in Music, comes down to finding patterns — hence the name. This is my quest for the patterns — in Materials and in Music.
Here you'll find materials datasets (pre/post-featurization) and results from individual projects — model and ensemble performance, feature engineering, optimization — across DFT, MD, and ML stages of the pipeline. Implementation details live on my GitHub. Downloads are provided to play with RDBMS tooling; please credit original sources for raw data.