friscofilth.com was my first major project, and it’s a site i use every day. i've been building a lot in 2026 (but haven’t been posting everything), and i've become a better developer. i decided to put some of that experience into rebuilding the back-end data pipeline.
now it picks up more details, like supporting acts, separate doors and show times, prices, and age restrictions. it tracks where those details came from so it can update existing shows automatically without creating duplicates or overwriting manual corrections. missing information stays missing, and conflicting details get flagged for review. i'm also now using typesafe's jev model to make classification decisions.
i tightened up the genre classification, too. a band name usually isn't enough to identify an artist, especially in this genre, so the pipeline pays closer attention to the context of the calendar event before making a judgment. uncertain shows go to review instead of getting published on a guess. it's also easier to spot when a venue's scraper breaks instead of quietly assuming there are no shows.
the south bay is now represented, with two venues added in san jose. as a bonus, i applied some new frontend skills i've learned and gave it more of a grimy aesthetic. there are better region and venue filters, more useful show details, and a “recently added” sort to see what’s new. less time checking calendars, more time going to shows.