A reliable bulk density (BD) data set is critical to the robustness of any Mineral Resource estimate (MRE) as it underpins tonnage and metal calculations. Representative coverage of BD measurements across a deposit can be difficult to achieve, as can consistency of measurement protocols. This is often the case when many phases of work have been completed, often by different companies. Machine learning (ML) has been used at the Dugald River zinc-lead-silver mine to generate additional proxy density data and to review historical measurements for consistency. The process has been refined over several iterations, and a hybrid data set is now used for Mineral Resource estimation.