<p>Cybersecurity threats have become increasingly complex, necessitating more robust detection and defense strategies. Cybersecurity data science uses data science and machine learning to tackle these challenges. This paper thoroughly examines cybersecurity data science, with a focus on machine learning applications. We discuss data collection, pre-processing, feature extraction, and machine learning implementation in the context of cybersecurity. We review top clustering, classification, and anomaly detection methods in the field. To ensure transparent decision-making, our approach [HYDRA (Hybrid Data-Driven Resilient Anomaly Framework)] includes model training and evaluation. We assess outcomes with metrics such as accuracy, precision, recall, and F1-score. Our results show that the proposed approach achieves 95% precision in identifying risks, outperforming conventional methods. This research supports SDG9.</p>