Analytics capstone for an industrial distributor. Built a sponsor-deployed system using behavioral clustering, churn modeling, and forecasting to guide sales prioritization.
Transformed 30 years of fragmented Excel records into a normalized PostgreSQL database with a Python ETL pipeline and executive dashboards in Metabase.
Kaggle competition predicting display ad CTR through iterative modeling from linear baselines to GAM. Ranked 18 out of 387, top 6%.
Built a data engineering pipeline processing millions of NYC parking records from PySpark ingestion to MongoDB storage to an interactive Streamlit dashboard.
Two-phase market research study using in-depth interviews and a survey experiment to identify which design cues drive purchase intent for Kirkland's online channel.
Designed a randomized controlled trial to identify which messaging strategies increase dog adoption rates, with full power analysis and causal inference framework.
Post-acquisition analytics on player engagement, sentiment, and monetization signals after Microsoft's $68.7B deal. Delivered as Tableau dashboards and an executive strategic proposal.
Analyzed 1.8M credit card transactions using unsupervised clustering and supervised classification including Random Forest, Neural Networks, and Deep Learning.