Education Research Using Data Mining and Machine Learning with Computer Science Undergraduates

In recent decades, we are witness to an explosion of technology use and integration of everyday life. The engine of technology application in every aspect of life is Computer Science (CS). Appropriate CS education to fulfill the demand from the workforce for graduates is a broad and challenging problem facing many universities. Research into this ‘supply–chain’ problem is a central focus of CS education research. As of late, Educational Data Mining (EDM) emerges as an area connecting CS education research with the goal to help students stay in their program, improve performance in their program, and graduate with a degree. We contribute to this work with several research studies and future work focusing on CS undergraduate students relating to their program success and course performance analyzed through the lens of data mining. We perform research into student success predictors beyond diversity and gender. We examine student behaviors in course load and completion. We study workforce readiness with creation of a new teaching strategy, its deployment in the classroom, and the analysis shows us relevant Software Engineering (SE) topics for computing jobs. We look at cognitive learning in the beginning CS course its relations to course performance. We use decision trees in machine learning algorithms to predict student success or failure of CS core courses using performance and semester span of core curriculum. These research areas refine pathways for CS course sequencing to improve retention, reduce time-to–graduation, and increase success in the work field