I’m having so much fun playing with Raleigh’s tree dataset and giving the trees something to say. Today I gave the trees some hyperlocal news mojo.

How it works: when you ask a tree for local news, the program checks to find streets nearby. It then uses a news API to check WRAL, WTVD, and WNCN for mentions of those streets in the last 90 days. Finally, it bundles that information into a ChatGPT prompt and generates a summary (while links to the news articles used for the summary appear beneath.)

Obviously how much the trees have to report depends on where they are. As you might imagine, a tree near the NCSU campus has plenty of gossip to pass along!

Up next: getting restaurant recommendations and local construction information from the tree. 😂