Foundry: Machine-Learning Ready Materials Science and Chemistry Data

MLMI2-CSSI developed Foundry to provide a collection of machine-learning ready datasets specifically for materials science and chemistry. Updated in January 2026, the repository serves as a discovery hub to bridge the gap between raw scientific data and model-ready inputs.

Use Cases

• Predicting material properties using chemical structure features
• Training regression models on materials science datasets for discovery
• Benchmarking chemistry-specific machine learning models against standardized inputs

Strengths

• MIT License allows for broad academic and commercial reuse
• Domain-specific focus on materials science and chemistry
• Maintained by the MLMI2-CSSI organization

Limitations

• Metadata does not specify total row counts or specific column headers
• Absence of sample data files in the primary repository index prevents immediate validation

Provenance

Source

MLMI2-CSSI

Freshness

Last updated January 2026.