Raw motion data generated by sensors affixed to a user performing exercises or poses in response to perceiving content annotated with a motion track and presented to the user, the motion track generated using a posture dictionary is collected. The exercises or poses are determined from the raw motion data using the posture dictionary. Motion data indicating the exercises or poses within the raw motion data is generated using the raw motion data. The user is graded in performing the exercises or poses by comparing the motion data with the motion track. The content is presented to the user based on the grading of the user.