Ryan McCurry

Data Science & Analytics Professional

Master's candidate in Applied Data Science at the University of Michigan (expected Spring 2027) — building on a background in SaaS and a foundation in Management Information Systems and Operations Management from the University of Arizona.

LinkedIn Profile ryan24mccurry@gmail.com

Background & Experience

Data scientist and analyst with a background in SaaS and experience translating complex data into actionable insights for cross-functional teams. Currently pursuing a master's degree in Applied Data Science (AI concentration, 4.0 GPA) at the University of Michigan. Full work history, education, and skills on my resume page.

View Resume →

Projects & Coursework

Technical Communication & Presentation

Resume Screening AI: A Data Ethics Analysis A video presentation analyzing the ethics of a resume screening AI, examining data provenance and power imbalance risks through frameworks from Onuoha (2016) and Kalluri (2020). Employee Turnover Prediction & Strategic Analysis A logistic regression model predicting employee attrition, paired with executive-level communication of findings and a ₹318M strategic recommendation.

Data Analysis & Visualization

NBA Shot Analysis: The Evolution of Modern Basketball An interactive Plotly dashboard analyzing 828K shot attempts across 14 NBA seasons, exploring how the three-point revolution reshaped shot selection and shooting efficiency. Fitness Performance Analysis: What GPS Data Reveals About Exercise A four-part visual analysis of personal GPS fitness data across 64 sessions, building from individual patterns to an interactive geospatial map. Gender Representation in Film: Implementing the APT Algorithm An implementation of Mackinlay's APT algorithm for automatic encoding selection, applied to Bechdel Test results across the top 20 grossing films of 2016. Building a Generalized Horizon Chart Renderer in Altair A from-scratch implementation of mirrored horizon charts in Altair, supporting positive, negative, and mixed time series with configurable bands and color palettes.

Machine Learning

Regression & Classification Fundamentals An exploration of polynomial regression, Lasso regularization, and the bias-variance tradeoff alongside KNN and SVC classifiers on synthetic and real medical datasets. Spam Email Classification: Precision, Recall, and the Cost of Being Wrong A binary spam classifier built with Logistic Regression and SVC, evaluated through precision-recall curves, ROC curves, and GridSearchCV — with a focus on asymmetric error costs. Classifying Physical Activity from Wearable Sensor Data A four-class activity classifier trained on 4,480 physiological sensor observations across 40 subjects, achieving 0.913 ROC-AUC with only 10 of 533 features.