0→1 Data Science. From applications to analysis.
AI is changing how we build software. I'm exploring what this means for full-stack data science workflows. I love what I get to do for work - leading data science teams, projects that reach millions of customers and influence real business decisions. I'm also too curious about what's coming next, so I explore in my free time.
I built this learning lab to practice the full 0→1 loop—prototype an app, wire up analytics, watch real users, analyze the data, and publish what I learn. The stack is Astro (static-first with React islands), PostHog (analytics + feature flags), Supabase (Postgres + real-time API), and Python notebooks for statistical analysis—all wired together with automated pipelines, to make it easier to build more web apps on top.
Everything is open source and non-commercial. Check out a live experiment or the analysis. If you learn by building too, feel free to follow along or contribute.
Projects
Each app applies a data science method to a real problem. Interact, contribute data, and explore the published analysis.
A/B Testing Memory Game
LiveA live A/B test studying how game difficulty affects player behavior. Play the memory game, contribute data, and explore the published analysis.
Recent Posts
- Technical Post
Sign up for my Newsletter
Updates from 0→1 data science. What works, what breaks, what I'm learning.