Visual Learning Analytics of Multidimensional Student Behavior in Self-regulated Learning

Abstract

In Self-Regulated Learning (SLR), the lack of a predefined, formal learning trajectory makes it more challenging to assess students’ progress (e.g. by comparing it to specific baselines) and to offer relevant feedback and scaffolding when appropriate. In this paper we describe a Visual Learning Analytics (VLA) solution for exploring students’ datasets collected in a Web-Based Learning Environment (WBLE). We employ mining techniques for the analysis of multidimensional data, such as t-SNE and clustering, in an exploratory study for identifying patterns of students with similar study behavior and interests. An example use case is presented as evidence of the effectiveness of our proposed method, with a dataset of learning behaviors of 6423 students who used an online study tool during 18 months.