Society is relying more and more on computer-generated information due to the online abilities provided by current information and telecommunication technologies in a variety of ways such as social networks, learning systems, shopping, quality-of-life improvements. Multimodal Learning Analytics (MMLA) is a method in learning analytics research that makes it possible to capture large amounts of data on human activity. This study aims to provide a deeper understanding of physical movement challenges for training performers in open-ended, practice based learning settings. Moreover, it discusses how multi-modal analytics systems can provide support for performers. This study identifies ten important requirements that a prototype should have in order to fulfill the performer’s needs. These requirements are implemented in a low fidelity prototype that provides modeling movement followed by Laban Movement Analysis theory, capture user data with MMLA tools, and provide personalized feedback. The results indicated there is a potential within the usage of a multi-modal system to support and improve the motor skill learning process through personalized help and feedback.