Women in Computing: Data-Driven Retention Analysis

Project Overview

This repository contains research analyzing the retention patterns of women in computing programs at BMCC (Borough of Manhattan Community College). Using machine learning techniques on student course data spanning multiple semesters, this project identifies key factors that influence women's persistence in computing education.

Key Findings

• Women in computing programs show different retention patterns compared to male counterparts
• Significant differences in course performance exist between genders in specific gateway courses
• Early course success strongly predicts retention through the computing program
• Cluster analysis reveals demographic factors (first-generation status, financial need, enrollment type) interact with gender to influence success outcomes
• Machine learning models can identify at-risk students based on early performance indicators

Methodology

This research employs a comprehensive data science approach:

1. Data Collection: Analysis of BMCC student course data across four semesters (2021-2023)
2. Cohort Tracking: Longitudinal tracking of student persistence through computing programs
3. Statistical Analysis: Identifying significant performance gaps and retention trends
4. Machine Learning: Building predictive models to identify key factors in student success
5. Interpretable AI: Using SHAP analysis to explain model predictions and identify success factors

Technologies Used

• Python: Primary programming language
• Pandas/NumPy: Data manipulation and numerical analysis
• Scikit-learn: Machine learning models including Gradient Boosting Regression
• Matplotlib/Seaborn: Data visualization
• SHAP: Model interpretability and feature importance analysis

Recognition

This research won Best Undergraduate Research Poster at the 29th Annual Consortium for Computing Sciences in Colleges Northeast Region (CCSCNE 2025) Conference.

Future Directions

Findings from this research are being used to develop the BMCC Women in Computing portal, designed to support female students through:

• Targeted mentorship programs focusing on identified gateway courses
• Skill development tracking based on predictive success factors
• Community building interventions to improve retention rates