make robot to score every single tag in #hoodmaps db (about 50,000+) by sentiment, then we can combien that with good/bad places, and the colorings and essentially build a heat map of good and bad in a city
create a backend LLM-based data analyzer to find Recommended Areas and Sketch Areas, it can use the tags written by people + drawing colorings + bad type places (like gambling, pawn shops etc.) to figure this out and update monthly #hoodmaps