Conventional logistics practices frequently utilize simplified metrics, notably mean container dwell times, which often obscure the intricate dynamics and fluctuations inherent in complex operations. This study proposes an innovative, data-centric methodology for scrutinizing temporal operational data within container logistics, specifically designed to overcome the shortcomings of aggregated metrics. The framework integrates one-hot encoding for categorical attributes, discretization of continuous temporal variables, and distance-based clustering to analyze operational patterns. A pivotal advancement is the selective treatment of outliers; instead of their elimination, outliers are strategically retained and analyzed within the clustering process, facilitating the identification and isolation of anomalous operational occurrences. The methodology was developed and refined using empirical container dwell time records from the Port of Sines (Portugal), showcasing its capacity to reveal bimodal distributions and other non-uniform patterns that remain undetectable with average-centric analyses. To confirm its adaptability and robustness, the methodology was also evaluated using simulated drayage operation data, where it effectively differentiated between standard operations, subtle deviations, and intentionally introduced outliers. High-Density Region (HDR) analysis further enriched the interpretability of the findings, supplying probabilistic intervals for dwell times. This research underscores the substantial benefits of transitioning from aggregate statistics to a distribution-aware, context-sensitive, and outlier-adaptive approach for analyzing complex operational data, thereby enabling more informed decision-making, refined resource allocation, and enhanced operational efficiency in container logistics.