Abstract
High costs are incurred for the collection and analysis of samples required for mine material geochemical characterization studies. The associated datasets, which are valuable and irreplaceable, offer a unique opportunity to explore advanced data analytical techniques such as machine learning that could support future innovation in the management of water and materials across the project life cycle. However, this project highlights systemic issues in data reporting that limit potential data reuse. This project independently leverages publicly available static test, kinetic test, and mineralogy data from impact assessment reports for metal mining projects across Canada since the 1990s. Specialized Python programming scripts, coupled with built-in manual verification checkpoints, were developed to extract, clean, and standardize datasets into a consistent format and nomenclature for inclusion in an aggregate database. This data transformation process relies on machine-readable data, accompanied by accurate and robust sample and analytical information or metadata. This paper highlights common gaps in the application of the Mine Environment Neutral Drainage 1.20.1 guidance document related to the documentation of sample origin and acid base accounting test methods. These issues are based on observations from 60 projects with accessible mine rock data, encompassing 28,507 unique mine rock samples, 20 sample metadata fields and approximately 125 parameters. Issues include data in non-machine-readable formats, missing sample lithology and location metadata, non-unique sample identifiers, non-repeatable test methods descriptions, and inconsistent provision of laboratory certificates of analysis. These issues erode data integrity, ultimately calling into question the reliability of the aggregated database contents and limiting data interoperability and reuse for innovative applications. This initiative demonstrates that independently leveraging data is not sustainable without the cooperation of proponents, and could be resolved through community-endorsed standards and minimum information requirements for study reporting.