Purpose (North Star). Give the Board of a regional U.S. health plan the evidence it needs to redesign how members are invited to its cardiovascular prevention program: replace single-criterion invitations (an age cut-off or a single diagnosis) with a combined risk score, so that limited cardiology capacity finds the most cases. Big idea. With the same number of invitations, a combined risk score finds up to 16 percentage points more heart-disease cases than any single-criterion rule. The Board should approve a 12-month staged pilot to confirm this in the plan’s own population.

Dashboard en Salud

An end to end R pipeline modeling the relationship between COVID-19 case surges and national diagnostic device shortages (e.g., sodium citrate blood collection tubes). Combines JHU CSSE epidemiological time-series data with FDA medical device shortage data acquired via web scraping (rvest/httr) to overcome bot-blocking. Applies cross correlation (CCF) time-series analysis to identify a 5-day lag between case spikes and official shortage reporting insight informed by my background managing coagulation instrument supply chains in clinical laboratories during the pandemic. Built using R (tidyverse, rvest, httr, lubridate) following a reproducible data pipeline architecture (Obtain → Scrub → Explore → Model → Interpret, OSEMN framework).