A framework that isolates meteorological drivers of PM2.5 concentrations. LightGBM and XGBoost models significantly outperformed OLS baselines in predicting urban air quality.

Automated ETL pipeline integrating the Scryfall API to normalize real-time Magic: The Gathering pricing data. Trained XGBoost/LightGBM regression models on playability signals and Reserved List status to detect emerging market trends.

A prediction engine that scrapes live lineups and match results, runs them through an ML model and 10,000-iteration Monte Carlo simulation, and outputs win probabilities, scoreline distributions, and a full bracket forecast for all 48 teams.

A Python-based financial database cleaning and management system developed during an internship at Indorama Ventures in summer 2025.

Implements an n-gram Markov Chain model in Python to generate synthetic text based on the Harry Potter series by analyzing word sequence frequencies and building state-transition probability matrices to stochastically produce new sentences that mimic the author's stylistic patterns.

A comprehensive data analysis and machine learning workflow built in Python/Jupyter Notebook that analyzes the evolution of college-level writing through data normalization, feature engineering, and integrated ML models to help faculty refine course design and students enhance their writing skills.