SQLearn: Automated SQL Statement Assessment using Structure-based Analysis

In the world of database education, SQL (Structured Query Language) is often the first and crucial step toward developing data analysis and manipulation skills. As the world moves on to data-driven technologies and businesses, the need to educate students in writing accurate and efficient SQL queries has become paramount. Traditional SQL evaluation modes often rely on limited, subjective, and labor-intensive manual grading, which impedes the integration of practical assignments into the curriculum. This poster introduces SQLearn, an innovative automated assessment tool for SQL education. We aim to build a comprehensive platform that addresses submission, evaluation, and review needs amongst students and educators. Here we highlight our assessment approach which breaks down student-submitted queries into Abstract Syntax Trees (AST) and uses cosine similarity to evaluate them. Experimental results show that the proposed approach is effective, not only in binary grading of queries but also in assigning partial grades. The tool also offers an interactive platform to submit and receive feedback, enabling students to refine their SQL skills and gain insights into query structure and optimization. By automating the assessment process, educators can focus on refining the curriculum and channel more time into instruction and research.