Data Science · Coffee · Curiosity
I'm Warren — undergraduate Data Science student at the University of Arkansas, former Navy diver, and a hobby-collector who finally found one that gives back. I build data visualizations, write statistical analyses, and ship browser-based machine learning demos.
My career started underwater. I spent nine years as a U.S. Navy diver out of Pearl Harbor — ship's husbandry, salvage and recovery from Southeast Asia to the Middle East, and emergency medical response for 7th fleet. Then nine more in electrical and electronic signals work, running squadron-wide casualty data collection and managing teams of technicians.
After the Navy I worked as a senior field applications specialist for Lutron, then ran my own business technology consultancy in Los Angeles. Along the way I picked up a data analytics fundamentals from a bootcamp at USC, kept coding through Pasadena City College, and shipped production tooling at KLA.
Today I'm an undergraduate Data Science student at the University of Arkansas — an interdisciplinary program across the Business, Science, and Engineering colleges. I program in Python, JavaScript, and C++, with a focus on agricultural commodity data, statistical analysis, and turning rich datasets into things you can actually click on. Lately that's been coffee. A lot of coffee.
Hobbies, by way of explanation: ♆ diving, 🪂 skydiving, 🏄 surfing, 🏂 snowboarding, 🧘 yoga, 💃🏽🕺 salsa & swing, and — for the long game — fermenting kombucha in a glass dispenser on my counter.
Hand-picked, not exhaustive. Click any card for the live demo; "Code" for the repo.
A five-chapter, browser-native rebuild of a Tableau workbook on global coffee — production, climate stress, who exports to whom, and where the United States fits in. Pure HTML, CSS, and JavaScript. No build step.
Do climate, population, and oil price predict coffee bean production and export revenue? Thirty years, ~60 producing countries, FAO + BACI/CEPII data. Published as an interactive Marimo notebook so every chart and statistic is reproducible in the browser.
An interactive map of recent earthquake activity from the live USGS GeoJSON feed, with a scrubbable timeline so you can watch seismic events populate in sequence. Built on Leaflet with a modern muted basemap and clean Google-Maps-style info cards.
Train an image classifier on your own webcam captures, then play Rock-Paper-Scissors against the model you just trained. The gamify-your-classifier idea I started in 2019 — currently being finished as a browser-only sibling to Hand_AI.
A webcam sticker board you control with your hands. MediaPipe tracks your hand in the browser; pinch to drag a sticker, point to spawn one, fist to delete, thumbs up to swap the emoji. Runs entirely on-device.
A full-stack juice-bar ordering application themed on Jabberwocky. Live SQL-Server-backed menu, cart, checkout, and order history. Shipped with a four-person team — I led app development and deployment.
A featured long-form piece. More on Medium.