The intended goal of this research is to enhance the safety surveillance of railway systems by implementing clustering and outlier identification algorithms. Unsupervised machine learning methods can be employed to organize a vast amount of sensor, operational, and surveillance data into significant patterns that demonstrate standard system behavior. This process also detects anomalies that may indicate potential hazards, such as equipment malfunctions, track defects, or perilous human actions. Outlier identification methods identifie anomalous occurrences that deviate significantly from expected patterns, whereas clustering techniques disclose the inherent structures within complex datasets without prior labeling. The integration of a variety of technologies enables real-time monitoring, data-driven decision-making, and early warning systems, thereby improving operational reliability and reducing the likelihood of accidents in railway systems.