The microstructural analysis of bainitic steels commonly relies on the experience of an expert. Given the complexity of bainitic microstructures, interpretations on the microstructure and its constituents can be manifold. This causes difficulties to extract reliable microStructure-Property (P-S) linkages of the material. This study proposes an open-source data-driven framework, providing trained and tested models for detecting and segmenting sub-structures in bainite, and ML-based classification models for identifying S-P linkages.