Modern railway marshalling yards use IoT sensors to automate control and operational data collection. At the same time, analytical systems are actively being developed that identify vulnerabilities in technological processes and find economically unprofitable solutions for managing them. However, the disparity of information storage formats in different systems makes it difficult to analyze, reducing the reliability of results and increasing processing costs. This paper proposes a hybrid telemetry conversion method that combines the use of a digital model of the station infrastructure and data logical interpretation. The method allows transforming low-level information by generating events related to technological processes, and a logical interpretation is a set of formal rules that describe context-sensitive conclusions based on existing data.. The developed approach ensures the formation of a unified analytical model, eliminating the problem of incompleteness and initial information inconsistency. Additionally, existing solutions are based on a reactive approach that allows data to be presented in a chronologically ordered stream, effectively identifying relationships between operations and minimizing costs when scaling systems. Implemented algorithms demonstrate the possibility of functional and adaptive expansion of analytical tools without significantly modifying the basic structure. Over 10 million records of railway wagon movements were processed. When analyzing average downtime at stations, the method revealed 22% anomalies, which, if eliminated, led to increased downtime, reduced fluctuations in correlation, and improved data consistency.