This project presents a comprehensive approach to anomaly detection in industrial screwdriving processes using machine learning (ML) algorithms. The primary hypothesis postulates that ML, applied to various process data from the screwdriving operation, can proficiently identify system performance irregularities, serving as a dependable anomaly detection instrument. The project follows several stages, beginning with data collection. A dataset consisting of time-series sensor data from the screwdriver, data from a UR10 robot, and audio data from a microphone was amassed to capture the aspects of the screwdriving process. Following this, preprocessing was employed to clean and format the data for further analysis. Subsequently, feature selection techniques were employed to identify the most informative attributes from the data, strengthening the predictive power of the ML models. The final step was the model-building phase, where classification algorithms were devised to distinguish between different classes of screws. This research is a collaborative effort carried out at Aalborg University, with a partnership from VELUX. They provided a real-world manufacturing context where this anomaly detection approach was derived from. In the future the aim is to enhance the quality and efficiency of their automated screwdriving production line by minimizing manufacturing anomalies. The successful execution of this project would validate the viability of integrating ML into industrial applications, but also sets the stage for future research focused on optimizing the performance of the algorithm and broadening its applicability to other manufacturing sectors.