From Map Nodes to YouTube Clips — A Self-Updating Skatepark Intelligence System
An always-on LLM system that discovers, enriches, and curates global skatepark data—automatically.
Project Details
- My Role
- I led everything end-to-end—from designing the system to coding the LLM processes and deploying the dashboard.
- Duration
- 1 week
- Year
- 2026
The Brief
Manually hunting down skatepark info was draining—static lists, no fresh content, and all on me. I needed a system that would run while I slept, surfacing every new skatepark and enriching the data on its own.
Approach
I architected a daily pipeline—using Cursor, I sequenced each LLM task into smart time slots. We rolled out a system that wakes up, discovers, enriches, and curates—all in perfect daily rhythm.
Outcomes
The system delivered clean, up-to-date skatepark data daily, cutting my hands-on time to zero. Beyond that, it became a blueprint—I can now replicate this approach for any research task I dream up.
Services
Tech Stack
Craft Notes
I discovered deep search was the key—going beyond OSM data, I enriched insights with Google’s deep results. Now, I’m scaling future tasks with Maestro.
Gallery
