Using Data Mining to Support Teacher's Intervention in a Robotics Class

Following students' progress in robotics classes is difficult because student groups usually take different and unpredictable paths in problem solving. To improve teachers' possibilities for intervention at the right time, Open Monitoring Environment allows the teacher to monitor and model the learning process based on the data rising from the current learning setting. Since the amount of data that can be extracted from the learning process is too large and complex for manual processing, there is a need for semi-automated tools to support the reasoning. Various data mining methods were tested with authentic data that was collected from a robotics class in a primary school. In the novel monitoring environment, initial rules produced by the classification algorithm are open for revision by the teacher. We present a proof of concept implementation for the data mining module in the Open Monitoring Environment. Results from the study indicate that decision trees are effective for classifying students' progress in the educational robotics setting, and that open data mining process produces useful and interpretable information about the students' progress with relatively small datasets.