FULL-STACK APPS

Two live platforms, built end to end.

Data Engineer · Full-Stack Developer · Project Manager

The data, the pipeline, the maps, the publishing — one pair of hands. Both products run on the same platform: federal data through Python into Azure SQL, out through the WordPress REST API.

CENSUS · FCCPYTHON ETLAZURE SQLWORDPRESS APICDN

CensusFlow

censusflow.com
live WebGL commute map — one dot ≈ 75 workers

Rebuilt a stalled one-city prototype into a nationwide commute & demographics platform — 50 states + DC, ~1,160 city pages — in 2.5 weeks.

~1,160
city pages
546M
rows in the OD table
7.5M
cross-state commuters counted
328M
people on the dot map
  • Engineered the full Python pipeline ingesting Census LEHD LODES (25 tables, ~546M-row OD table) into an Azure SQL warehouse, transforming it into city data marts and publishing thousands of static pages via the WordPress REST API with Azure Blob Storage as CDN.
  • Implemented cross-state commute attribution from scratch through a 100M+ row home-block join — correctly counting the ~7.5M Americans who commute across state lines (NJ→NYC, VA/MD→DC).
  • Unified four federal sources per city: LODES, ACS 5-Year (income, age, education, poverty, race via Table B03002), 2020 Census P.L. 94-171 block race counts (8M+ blocks, 328M people) powering an automated racial dot map, and FCC broadband across 1,732 cities.
  • Built the front end in ~2,000 lines of vanilla JS with two WebGL maps (deck.gl + MapLibre GL) rendering commute and racial dot-density visualizations.
PythonAzure SQLdeck.glMapLibre GLWordPress REST APIAzure Blob CDN

FindBetterInternet

findbetterinternet.com
H3 hexagon coverage map — speed-colored, population-weighted

A platform that ranks home internet by real availability at a given address — measured coverage, not commercial placement.

105
providers onboarded
~800
city pages
51
state guides
~80
providers scraped monthly
  • Acquired a preexisting domain and onboarded 105 providers (AT&T, T-Mobile, Spectrum, Verizon, Xfinity, Starlink and more) through the Red Ventures affiliate program.
  • Designed a 0–100 Speed Score that percentile-ranks census-block FCC serviceability data, population-weighted with Census LODES — so measured coverage beats advertised speeds.
  • Engineered the Python pipeline on Azure SQL and published 100+ providers, 34 reviews, 51 state guides and ~800 city pages via the WordPress REST API.
  • Built an AI plan scraper covering ~80 national and local providers — Claude API web search combined with Browserbase — feeding an ETL pipeline that refreshes prices, plans and promotions across the site monthly.
  • Built a MapLibre GL H3-hexagon coverage map and a custom SVG engine plotting population-scaled, speed-colored city markers — all feeding a monetized address-check funnel kept separate from the rankings.
PythonAzure SQLClaude APIBrowserbaseMapLibre GLH3Red Ventures