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

Classification using logistic regression and model selection criteria

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

Linear regression using different variable selection techniques

NPV, IRR, and capital budgeting

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

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

Diversification, efficient portfolios, capital market line, and CAPM

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

Latent-class count models using the NBD and Poisson

Modeling count data using the negative binomial distribution

This project was developed as part of the Wharton High School Data Science Competition(https://wsb.wharton.upenn.edu/wharton-data-competition/), organized by the Wharton Sports Analytics and Business Initiative. The competition provided teams with simulated soccer league data to develop predictive models for playoff outcomes.

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

Basketball Prediction System built for the 2025 Wharton Data Science competition