This article presents VISGRADER, an innovative method for automatically grading D3 data visualizations. Manual grading of such visualizations is challenging, especially in large classes. VISGRADER evaluates data bindings, visual encodings, interactions, and design specifications in a scalable and precise manner. It enhances the learning experience by allowing students to frequently submit their code and receive quick feedback, thereby facilitating improvements in their code and visualization design. The method has been successfully deployed and tested with over 4000 students in a visualization course at Georgia Tech.
Publication date: 23 Oct 2023
Project Page: xx.xxxx/TVCG.201x.xxxxxxx/
Paper: https://arxiv.org/pdf/2310.12347
