Augmenting Teachers with Data Science Powers: Joining Human and Artificial Intelligence to Assist Students

Ranilson Paiva, Júlio Ferreira Silva de Holanda, Myron David Peixoto, João Pedro Vieira

Resumo


Technology has the potential to influence the processes and outcomes of teaching and learning. Researchers notice a high demand for online education, but these environments generate educational data in high quantity, diversity and velocity, which requires the use of technology. However, putting too much emphasis on artificial intelligence is not bringing the expected results. It is time to position teachers as the principal decision-makers in the ``classroom'' and use AI as an assistant. Our proposal is a tool (called T-Partner) to help online teachers: (1) search for relevant patterns in the educational data; (2) created visualisations to present the results in a more teacher-friendly way; (3) recommend study plans, based on students' needs. For that, we applied data science techniques on these data and processed the results, making it easier for teachers to understand them. We evaluated the tool, asking participants to use it to complete some tasks that simulate a teacher analysing data to make pedagogical decisions. We evaluated time, correctness, number of doubts and the participants' perceptions. The results show that most participants were able to complete the tasks in an appropriate amount of time and considered the T-Partner useful and easy to use in an online learning context.

Texto completo:

PDF


DOI: https://doi.org/10.5753/cbie.sbie.2019.1721