Edinburgh Napier University

  In this paper, we present our findings of the design, development and deployment of a proof-of-concept dataset that captures some of the physiological, musicological, and psychological aspects of embodied musicking. After outlining the conceptual elements of this research, we explain the design of the dataset and the process of capturing the data. We then introduce two tests we used to evaluate the dataset: a) using data science techniques and b) a practice-based application in an AI-robot digital score. The results from these tests are conflicting: from a data science perspective the dataset could be considered garbage, but when applied to a real-world musicking situation performers reported it was transformative and felt to be ‘co-creative’. We discuss this duality and pose some important questions for future study. However, we feel that the datatset contains a set of relationships that are useful to explore in the creation of music.