The efficiency of planning processes is fundamental to the success of modern mining enterprises, particularly in maintaining competitiveness in the raw materials market. This paper addresses the critical challenge of optimising production lines, encompassing strategic arrangement of mining fronts and the management of drilling, transport, and auxiliary machinery. Given the dynamic and often challenging conditions of underground mining, mitigating random disruptions that can lead to machine downtime is crucial for sustaining productivity. This study highlights the importance of leveraging large volumes of diverse data to achieve situational awareness and facilitate informed decision-making under uncertainty. We explore how text mining and machine learning techniques can extract valuable insights from unstructured data, such as equipment failure reports, which often contain complex and ambiguous information. By developing a classification system that categorises failure descriptions into actionable insights, we aim to improve data interpretation and support predictive maintenance strategies. Furthermore, we propose an integrated approach that consolidates data from various sources, including downtime logs and repair records, to establish a comprehensive database for analysis. The proposed methods not only streamline data management but also enhance the accuracy of predictive models, ultimately enabling mining companies to optimise their operations and increase overall productivity.