The teacher in the loop: customizing MMLA in blended CSCL scenarios

In across-spaces learning scenarios, evidence needs to be gathered from different spaces to obtain a more complete view of the teaching and learning processes. Multimodal learning analytics (MMLA) enables us to gather data from physical spaces, enriching the evidence coming from digital ones. However, blended learning scenarios are heterogeneous, and the varying data sources available in each particular context can condition the accuracy, relevance, and actionability of the analyses. To avoid this problem, in this paper we propose to involve teachers in customizing the LA solution they will use, adapting it to their particular blended learning context (e.g., identifying relevant data sources and metrics). Preliminary results from two studies in blended CSCL settings show an improvement in the accuracy of the resulting MMLA solution. Although this kind of approach requires additional time from teachers, participants reported increased levels of relevance, novelty, understanding and actionability of the results.