Student Risk Model

Overview

Multi-stage optimization and logistic regression model to identify at-risk students

Key business outcome: Increased student success rate by 9% while reducing advisor case load by 32% vs. previous strategy. Increased student revenue while reducing advisor costs.

Skills

• Programming (Python, Docker)
• Statistical and numerical analysis (Pandas, NumPy)
• Machine learning (scikit-learn)
• Relational databases (Amazon Redshift/PostgreSQL)
• Cloud Computing and AWS Deployment (AWS SageMaker, Apache Airflow)

The Take-Away Message

Page in progress