Stefan R.

I find the data errors that quietly cost companies money: prices that don't match, totals that don't tie, and rules that calculate the wrong result. Most of my work has been in travel. As the sole data-validation analyst on a travel loyalty-optimisation platform, I independently QA'd a loyalty-and-pricing engine with 100,000+ rules, and I validate hundreds of fares daily. The same method works on any data you need to trust: take two sources that should agree, and report exactly where they don't. WHAT I DO - Data validation and reconciliation: compare a system's output against an independent source, row by row, and report every mismatch with its cause - Web scraping and price monitoring: collect prices from major travel booking sites, including sites with bot protection, with a quality check before anything is delivered - Python data pipelines and automated monitoring: API ingestion → PostgreSQL → analytics, run daily on a schedule, with price history kept between runs and alerts sent only after they pass validation - Independent calculation models that verify pricing, earning and status logic - Spreadsheet work: Excel and Google Sheets models, clean-up and automation HOW I WORK Nothing reaches you unchecked. Every dataset and every alert passes a QA step first, so you get clean, trustworthy data instead of false signals. TOOLS Python (pandas, SQLAlchemy, Playwright) · PostgreSQL / Supabase · SQL · GitHub Actions (scheduled pipelines) · flight-search & award-availability APIs · Excel · Google Sheets · Streamlit · Docker · Jira A recent client put it simply: "Stefan helped us set up and test our complex logic system with over 100,000 rules — no small task." If you have data you can't fully trust, tell me what's breaking and I'll tell you how I'd validate it.