In recent years, advances in Data Science have led to the investigation of various Structural Health Monitoring methods. This study aims to provide an innovative data-driven technique for automatically locating structural damage in bridges based on features extracted from dynamic data. The proposed method utilizes advanced multi-domain and filtering approaches to extract relevant features from sensor data in the temporal, frequency, and quefrency domains. By using a feature selection (FS) procedure, the method reduces redundancy and increases the relevance of the feature set, allowing more accurate damage localization. The proposed technique is versatile and can be customized to a specific structure while also being generic enough to handle any form, material, or excitation the structure receives. The study also introduces a damage index generated through an outlier analysis using the structure's healthy state as a reference. The proposed procedure has the potential to save time, cost, and resources by reducing the need for manual inspections, accelerating decision-making, and allowing timely maintenance. The methodology is validated in a full-scale bridge under various environmental conditions, such as traffic, temperature, and wind speed changes, showing promising results for real-world monitoring. The study aims to contribute to the ongoing development of efficient and cost-effective structural health monitoring methods in the field of data science.