Visualizing Learning Behavior as Spatio-Temporal Trajectories

Abstract

To tackle the proposed problem, several mature research streams can be considered. For this PhD project, especially CBR, schema matching for integration scenarios and component adaptation are currently regarded as relevant. The digitalization of teaching and learning has become an increasing desire for schools and universities. In order to apply digital media purposefully, educational organizations need to understand if and how students make use of digital contents and platforms. In the following we present a technique that uses arbitrary logging data as they may be present in any ICT systems that are commonly used to distribute digital learning contents. It transforms arbitrary data into spatio-temporal trajectories that can be analyzed only on the basis of their geometric relationships and characteristics. Through this we lift heterogeneous data to a highly abstract level. In an example, we illustrate how we can distinguish different types of users regarding temporal patterns and the learners’ mobility. We are also able to recognize groups of students working on similar topics. We mostly understand the current state of our system as a tool that can give both researchers and teachers the possibility to examine student’s behavior on a qualitative basis. In an outlook we furthermore describe how the system can be extended to support automatic clustering of learning behaviors.