The Spanish livestock sector plays a key role in European agricultural production, recently benefiting from investments in advanced technologies to optimize the management and sustainability of the sector. This research work focuses on the analysis of animal behavior, using smart monitoring collars to collect data from dairy cows in intensive and extensive farms. Using dimensional reduction techniques such as PCA, Isomap and t-SNE, this paper compares behavioral patterns between both kinds of farms. The main goal is to identify significant differences in behavior that may influence animal health and efficiency. The results indicate that these techniques can effectively discern between behaviors under different management regimes, which is essential for the implementation of future intelligent models to aid in the early detection of abnormal events such as disease or reproductive changes.he ability to characterize and differentiate these behaviors underscores the importance of advanced data analysis techniques in the continuous improvement of livestock management, aligning production with more sustainable and responsible practices.