Sorry for the stupid question but do you have a writeup or the code anywhere? I've always loved the idea of using an LLM to analyze massive amounts of raw data like you are with these tracks but idk how to format it in a way that allows the LLM to make sense of or analyze it