Data Analyst · Las Vegas, NV
I build interpretable, deployable ML systems for high-stakes industries. I've spent my career in environments where analytical errors have real consequences — a restaurant lives or dies on margins, an NGO allocates scarce resources under legal constraints, a sportsbook loses money to customers who are smarter than the line. That domain grounding distinguishes this portfolio from candidates who have credentials but not context.
Healthcare · Finance · Supply Chain · Database/NLP · Real Estate · Tourism · Agriculture · Insurance
17-table, research-grounded synthetic hospital database with a generate-validate-repair data quality pipeline — no PHI, built from peer-reviewed clinical/operational schemas.
Weak sentiment-return correlation turns out to be the interesting result — bot prevalence appears to suppress the organic sentiment signal in equity Twitter data.
ARIMA vs. LightGBM on retail demand — documenting where ARIMA's stationarity assumptions break on promotion- and holiday-driven sales data.
Text-to-SQL interface over a normalized analytical schema — plain-English questions translated to validated SQL via an LLM query layer.
Geospatial price model where flood-zone risk and coastal proximity are first-class features — VE zones expected to show a premium despite highest risk.
Multilingual (English + Spanish) sentiment and topic modeling across Miami/Orlando hotel reviews — reputation scoring and competitor benchmarking.
Weather-driven yield forecasting for Florida citrus or strawberry — frost days and rainfall expected to dominate feature importance.
Per-property Annual Expected Loss scoring from FEMA NFIP claims and NOAA storm tracks — Florida accounts for 79% of US homeowner insurance lawsuits.
This portfolio bridges a decade of experience in high-stakes operational environments — hospitality, humanitarian services, and financial risk — with a formal data science skill set built through an M.S. in Data Analytics and applied project work. Two years as an equity trader at William Hill / Caesars Sportsbook, including market analysis and data-driven presentations to leadership, is the most directly transferable professional data science experience and grounds several of the projects above, particularly the market sentiment and risk-modeling work.