Artifacts for Enhancing Software-Design Learning by Combining UML and LLM-Assisted Feedback in a Learning Platform

This artifact package contains the materials required to reproduce and inspect the implementation and evaluation of the proposed AI-assisted learning platform for software design. The repository includes: Frontend source code: The complete implementation of the web-based learning platform, including the user interface for learning modules, UML diagram creation, question solving, and feedback visualization. Backend source code: The server-side implementation responsible for user management, question delivery, submission processing, AI-assisted feedback generation, scoring, and database interactions. The prompt templates used to evaluate UML diagrams with Large Language Models (LLMs) are included within the backend repository under the prompts/ directory. Question dataset with evaluation criteria: The collection of software-design learning questions with their corresponding evaluation criteria used throughout the study, including their associated UML tasks and expected learning objectives. User study materials: The participant questionnaire administered during the evaluation, including demographic questions, System Usability Scale (SUS) inspired study-specific Likert-scale questions, and open-ended feedback questions. Anonymized user responses: The responses collected from participants during the user study, including questionnaire results and evaluation data. All personally identifiable information has been removed to preserve participant privacy. Together, these artifacts provide the implementation, datasets, evaluation materials, and collected study data necessary to understand, reproduce, and extend the research presented in the paper. The source code enables deployment of the learning platform, while the accompanying datasets and questionnaires support replication of the empirical evaluation.