Machine Learning HPDC datasets

This dataset contains 1 000 industrial heats of high-pressure die-cast (HPDC) Al-Si-based alloys, compiled to support the accompanying CALPHAD-oriented manuscript on machine-learning-driven optimisation of hardness (HB) and electrical conductivity (%IACS). Each record couples the melt chemistry, key thermodynamic descriptors and measured properties, enabling supervised learning, feature-importance analysis and inverse alloy design.