Jerold Crew

Data analytics senior (3.77 GPA, anticipated magna cum laude) specializing in SQL, data visualization, predictive modeling, and data governance. Performance Instructor at Lowe’s Distribution Center with experience training teams, creating onboarding materials, and improving safety compliance. Accepted into the University of Findlay’s MS in Applied Security & Analytics program to expand expertise in cybersecurity, governance, and risk management. Combines data-driven insights with collaborative leadership to resolve issues, improve processes, and support organizational decision-making.

Predicting Employee Attrition - August 2025

• Performed EDA and feature engineering on an HR dataset of 1,470 employees in Python, including variable encoding (labeling and one-hot), binning, and feature selection via chi-square, ANOVA, and correlation testing.
• Built logistic regression and random forest models; 98.6% accuracy validated through 3-fold cross-validation.

Storm Events and Crime Analysis – June 2025

• Analyzed Miami crime and storm records (2017–2019) in R to compare financial losses from crimes during storm versus non-storm conditions.
• Determined that storm-related crimes caused significantly higher victim losses (up to 35% increase during hurricane periods).

Gap Analysis for Sales & Inventory Strategy - April 2025

• Performed gap analysis on sales and inventory data, identifying seasonal sales peaks in September/November and dips in February/July.
• Found that office supplies significantly outperformed furniture and technology products, guiding recommendations for marketing and inventory alignment.

Heart Disease Risk Prediction - June 2024

• Developed logistic regression and random forest models in R to predict heart disease risk from 14 health variables.
• Compared logistic regression and random forest models; the second logistic regression model was chosen as the best predictor
• (76% accuracy, AUC = 0.85) due to stronger generalizability, while the random forest model showed overfitting (99.6% training vs. 68.9% testing accuracy).
• Identified chest pain type and exercise-induced angina as key predictors.

Housing Price Prediction Modeling - December 2023

• Built first-order, second-order, and nested regression models in R on 2,692 home sales, applying nested F-tests to weigh model
• complexity against predictive power.
• Found that living space increased home price by $129 per sq ft and lake views added $249K; selected a first-order regression model for interpretability (R2 = 0.81).