Materials for class on topics in deep learning (STAT 991, UPenn/Wharton)

Github Action wrapper for the Jake Wharton's Diffuse library

Github Action wrapper for the Jake Wharton's dependency-tree-diff tool

Shiny mobile app that has users select whether headlines are real or fake (from Buzzfeed survey); connects to mysql DB on EC2

Final project using a survey commissioned by Buzzfeed to identify the factors that best predict whether someone can correctly distinguish between accurate and inaccurate news headlines

A homework assignment from Seth Stephens-Davidowitz's class called Understanding Behavior with Big Data

Linear regression using different variable selection techniques

A lecture on the basics of exploratory data analysis using tidyverse as a TA for Wharton's Statistics Department's STAT701 - Modern Data Mining.

As a part of Business Analytics specialization from Wharton School, University of Pennsylvania, worked on the capstone project.

Applying NBD count models to examine the behavior of Wharton MBA students on the messaging platform GroupMe

NPV, IRR, and capital budgeting

Classification using logistic regression and model selection criteria

Diversification, efficient portfolios, capital market line, and CAPM

Latent-class count models using the NBD and Poisson

Timing models such as the exponential-gamma to measure time to purchase

Regularization using LASSO and ridge regression; cross-validation for parameter selection

Builds Google Sheet (600+ visits) with course and instructor evaluations and clearing prices distributed from the Wharton Analytics Club; visualization to help select courses

Assisted Wharton professors with data analysis for a paper on the changing of brand value over time

Brand concentration using count models; means and zeroes and method of moments estimation

Basketball Prediction System built for the 2025 Wharton Data Science competition