Human-Centred Teamwork Analytics
Objective 2
A novel methodology will be created to ground multimodal quantitative teamwork data in the semantics derived from a qualitative context interpretation. In a feasibility study, we proposed a conceptual representation for adding meaning to multimodal data, which will be used in this project (shown in grey in the figure).
Boxes represent artefacts such as data structures and interfaces.
Circles represent input parameters expressed as rules or templates to generate data structures or configure visual interfaces. Multimodal low-level activity data (D) are encoded into a meaningful information structure called the multimodal matrix (M) based on domain knowledge (K), including constructs from learning/teamwork theory and professional guidelines.
Each data modality can be modelled as one or more of the n columns of the matrix and associated with a teamwork construct or sub-constructs. Returning to the team leader’s positioning example, lines in the M matrix would contain meaningful positions each nurse was at during various moments of the simulation (e.g. the bedside or foot of the bed) instead of (x, y) coordinates.
In sum, realising this modelling structure would enable automated coding of quantitative, low-level data into qualitative, behavioural markers that can be grounded in generic features of all teamwork, as well as the specifics of the particular activity. This information can be directly visualised (V) but cannot be associated with the feedback communication referring to higher-order constructs (the focus of Objective 3).
