Software engineering education increasingly emphasizes hands-on experience through real-world projects. However, while technical proficiency is rigorously assessed, soft skills (such as communication, leadership, and adaptability) are often neglected, impacting team performance and project outcomes. This study aims to categorize student teams based on their soft skills profiles and predict their likelihood of success or challenges throughout the academic year. By identifying patterns in team compositions, we seek to anticipate potential difficulties and intervene proactively. Data from instructor evaluations of soft skills , weekly statuses of the projects, and the grades of the groups are analyzed to create team categories. These categories help forecast team dynamics and project success based on skill distribution and interpersonal factors. Early findings indicate that certain team compositions are more prone to issues such as leadership gaps or poor communication. Utilizing predictive modeling, we aim to enhance intervention strategies, ensuring more balanced and effective teams. This research provides a data-driven approach to understanding and improving team performance in software engineering education. The insights gained can guide team formation processes and support targeted educational interventions to maximize student success.