UC Riverside

Data science is a powerful tool for comprehending system behavior, identifying complications, and optimizing decision-making processes. Computer science education is a data-rich system ripe for exploration and understanding through data science methodologies. This data includes but is not limited to information from surveys, feedback systems, official academic records such as grades and DFW rates, registration records, and participation statistics. This study proposes a data-driven approach to enhance the undergraduate experience across various dimensions. We employ quantitative and qualitative methods to understand and improve three critical aspects of undergraduate CS education. This work addresses critical challenges in computer science education, particularly the development of communication, leadership, and teamwork skills among undergraduates. The COVID-19 pandemic, coupled with a significant surge in enrollment in computer science and related programs, has exacerbated these issues, highlighting the need for effective communication training and robust academic support for students navigating foundational coursework. First, we present an oral communication intervention integrated into a capstone course, designed to help students practice presentation skills and receive constructive feedback from professional facilitators. Next, we introduce two major frameworks that provide students with hands-on leadership and teaching opportunities alongside technical skill development: The Data Science Academy (DSA) trains undergraduates to lead outreach coding camps in data science for K-12 students, offering leadership and teaching experience in an in-demand field. The Undergraduate Learning Assistants (ULA) Program hires and trains undergraduates to tutor their peers in foundational computer science courses, addressing the need for accessible, timely academic support in large classes. We evaluate the impact of these programs on both the participants and the undergraduate leaders, providing a replicable model for other institutions. Finally, we explore strategies to enhance student performance in foundational courses by promoting positive academic behaviors and identifying areas of concern. Specifically, we study the effects of incentivizing early engagement with lab assignments in “Machine Organization and Assembly Language Programming”, a core computer science course. Additionally, we introduce a novel outlier detection approach to identify students at risk of academic misconduct or in need of support, using midterm exam and lab performance data. Together, these initiatives bridge gaps in professional skill development, academic support, and integrity, offering actionable solutions to enhance the undergraduate computer science experience.