Work Samples
Accenture.com Redesign
Wendy's Canada Breakfast Menu Launch
End-to-end Access Customer Journey - Info360 Asset
Autodesk Kudos
Telco Churn Analysis Presentation
R Code
For my Telco Customer Churn Analysis for my final project in Data Science for Business at Goizueta Business School, I explored why customers leave and built data models to predict churn with about 80–85% accuracy. Using logistic regression, Ridge and Lasso regularization, and a deep neural network, I identified key drivers such as contract length, internet service type, billing method, and demographics. For example, customers on month-to-month contracts or paying by electronic check were far more likely to churn, while long-term contracts reduced churn risk significantly. These insights translated into clear business strategies like incentivizing longer contracts, promoting paper billing, and investigating fiber optic service issues. This project demonstrates my ability to clean and prepare data, apply multiple modeling approaches, evaluate their performance with metrics like AUC and ROC curves, and most importantly, turn technical results into actionable recommendations for the business.
Walmart SuperForecast Presentation
R Code
For my Walmart Sales Forecasting final project for SuperForecasting at Goizueta Business School, I built a deep learning model to predict weekly sales across 45 stores using economic, holiday, weather, and store data. I chose a Long Short-Term Memory (LSTM) network because it captures temporal sales patterns, seasonality, and nonlinear relationships that simpler models miss. By merging and cleaning multiple datasets, creating lag features, normalizing values, and adding embedding layers for store and department context, the model achieved strong performance overall—explaining about 68% of sales variation across all stores, and over 95% accuracy for top-performing locations. These forecasts support inventory planning, staffing, and promotions, while highlighting that low-volume or irregular stores require additional external data to improve accuracy. Ultimately, this created clear business value by enabling Walmart to reduce waste, improve operational efficiency, and align marketing efforts with predicted sales trends.
Prototypes
To interact with the XLearning prototype above please follow the steps below:
1. Click "Get Started"
2. Click "Racial & Ethnic Equality
3. Click "Launch Simulation"
Accenture XR Ideathon 2021 Presentation
Graphic Designs
Animations
