Uche Nwoke, Brenda Ulin & Grant Brown, University of Iowa
Predicting student success is an important tool for educational institutions; however, such efforts are subject to competing requirements for student, collegiate and institutional specific objectives. Analytics must provide accurate student-level predictions, reliably forecast aggregate outcomes, and provide informative and actionable qualitative business intelligence. We discuss our collaborative approach to these issues, focusing on three components of data science that have been key to our success: data architecture, machine learning and statistical modeling, and integration of analytics into campus practice. We share strategies to navigate the complex balance between individual and aggregate objectives and the impact of identifying influencing factors.