Maintenance strategies and accurate claims analysis are crucial in transportation industries, where unplanned incidents can result in significant financial, operational, and environmental losses. However, despite the potential value of unstructured text data from maintenance logs, work orders, and claims data, they often remain underutilized due to processing complexities. This study presents a novel text-mining approach to analyze system degradations in large maintenance data. While we demonstrate this framework using industrial datasets that are too vast for manual analysis, the method is adaptable to diverse maintenance-related text sources, including claims, work orders, and maintenance logs. By applying advanced Natural Language Processing (NLP) techniques, we transform unstructured text data into structured actionable information about operational anomalies, their structure, and temporal trends. This enables organizations to implement proactive maintenance planning, optimize system design, and facilitate knowledge transfer. This framework bridges the gap between data volumes, data-driven approaches, and expert interpretation, empowering field specialists to derive actionable knowledge from complex datasets. By transforming raw text into interpretable formats, this approach establishes foundational insights for refining organizational policy and improving predictive maintenance models. Its versatility in analyzing diverse maintenance text data sources sets a new standard for proactive maintenance strategies across industries.