Predictive Modeling & Machine Learning
Amazon Electronics Sentiment Analysis
Developed an NLP pipeline in Python to analyze 1.1 million Amazon electronics reviews using NLTK and TextBlob, extracting sentiment trends and feature-level insights for e-commerce optimization.
Email Engagement Classifier
Built a Scikit-learn pipeline with StandardScaler and L1-regularized Logistic Regression (SAGA solver) on 40 engineered features from user logs to predict email click-throughs.
NFL Pass Completion Logistic Model
Implemented Logistic Regression on 8,704 NFL Big Data Bowl plays, selecting four key features via LassoCV to predict pass completions, and evaluated precision/recall.
Minnesota Twins Performance Explorer
Created an R Shiny app using Lahman database, applying PCA (FactoMineR) and GLM for HR prediction, and deploying interactive Plotly radar charts for player clustering and forecasts.
Spatial & Cultural Analytics
Wisconsin Census Geo-Analysis
Used GeoPandas, rasterio, and SciPy to mask NLCD rasters and train Ridge models predicting county population and visualized residuals via Folium choropleths.
Museum APIs: Islamic Art Study
Queried REST & GitHub APIs (Met, Louvre, V&A) in Python to analyze 18,000 acquisitions over 48 years, visualizing volume and metadata tag shifts with matplotlib.
Optimization & Simulation
Airport Baggage Carousel Scheduling
Developed JuMP MILP and two-stage decomposition in Julia to minimize passenger-weighted wait and bound max wait via a min-max model.