Data Analyst focused on statistically validated analysis — every notebook below runs end-to-end and gets checked against the underlying data before a finding ships. If a conclusion doesn't hold up, it doesn't go in the report.
Filter by discipline. Each project links to a fully executed notebook — real outputs, not narrated screenshots.
Does a new feature actually help the business, or does it just look good on paper? This project ranked which ideas to build first, then proved one of them lifted sales by nearly 19% — while catching a second "win" that turned out not to be real.
Which phone plan actually makes the company more money? This project rebuilt the billing math from scratch to find out — and caught an earlier assumption about which region performed best that turned out to be backwards.
Does bad weather really slow down rides, or does it just feel that way? This project proved it does — trips take 21% longer — giving a new rideshare company real numbers to plan pricing and driver availability around.
Where should the marketing budget actually go? This project traced revenue back to each ad channel and found one was quietly losing money while another was the real growth driver — redirecting spend toward what works.
When do customers come back, and when are they about to leave for good? This project mapped shopping habits across millions of orders to find the exact window to reach out before a customer churns.
Did a small design change actually change customer behavior, or was it a coincidence? This project tested it properly and confirmed the change was safe to ship — protecting the team from reading too much into random noise.
What actually makes a video game sell — great reviews, the platform, the region? This project compared sales across markets and genres to guide which titles to stock ahead of the holiday season.
Which customers are worth the most over time, and who deserves VIP treatment? This project built that picture from raw customer records using only core problem-solving, no pre-built shortcuts.
Messy, inconsistent customer data is unusable data. This project cleaned and standardized a real-world user database so it could actually be trusted for analysis.
Do people in different cities listen to music differently by day and time? This project compared two cities to find out — useful for deciding when and where to run ads.
How much is a used car really worth based on its mileage? This interactive tool lets anyone explore 50,000+ real listings to see how value drops off — and where the resale sweet spot sits.
Open to remote roles on US/Canada-aligned time zones. Let's talk.