1. INTRODUCTION AND SCOPE Efficient management of geochemical data represents one of the main technical challenges in the environmental assessment of mining projects, particularly in the analysis of long-term leaching tests. In this study, leaching devices (barrels) were used, composed of residual materials from mining operations, such as waste rock, tailings, and borrow material. These materials were configured under controlled treatment conditions to simulate real behavior and the chemical interactions among mine components: ● Control samples: materials without the addition of alkaline agents, used to establish a baseline behavior. ● Solid additive treatments: dosing of calcium carbonate (CaCO₃) in layers interbedded with the material. ● Liquid agent treatments: application of calcium oxide (CaO) as lime slurry in layered configurations. ● Combined systems: simultaneous use of solid and liquid additives (CaCO₃ and CaO) interlayered with the material. The diversity of experimental configurations generates extensive time series of physicochemical parameters and metal concentrations, which are critical for forecasting acid rock drainage (ARD) generation, evaluating metal leaching, and designing water management and mine closure plans. However, data dispersion, lack of standardization, and reliance on manual processes hinder comprehensive analysis. Addressing these limitations requires new approaches capable of extracting strategic value from large historical datasets that would otherwise remain underutilized due to the complexity of their processing and consistent structuring. In this context, the present work focuses on the development of a technology-driven solution based on artificial intelligence aimed at optimizing the processing and analysis of large volumes of geochemical data. The proposed approach emphasizes the automation of key stages within the data workflow, from data acquisition and validation to analysis and visualization, enabling a more consistent and timely interpretation of relevant geochemical processes. The scope of this research includes the integration of historical geochemical and hydrochemical datasets from multiple information sources, the implementation of automated ETL processes, and the development of analytical and predictive models for assessing the consumption and depletion of sulfides and neutralizing materials or additives. In addition, the approach prioritizes interactive data visualization, allowing specialists to explore scenarios, identify trends, and support decisionmaking for long-term environmental and operational management in mining projects. 2. METHODOLOGY The solution was structured following the Cross-Industry Standard Process for Data Mining (CRISP-DM) methodology, focusing on the integration of historical datasets, the automation of
RkJQdWJsaXNoZXIy MTM0Mzk2