New post on

Build Your Own Content Curation Pipeline

I've decided to take a different direction with Lagukan. Instead of focusing on a music recommender system, I'm building a general-purpose recommender system first (I'll drill into subdomains like music after that's established). The most important part of any recommender system is the data source. I've decided to use Atom/RSS feeds as the data source for a few reasons, one of which is that they're protocols, not platforms (<- great article I read after Twitter's big announcement today). I'm working on making it easier for people to create feeds of content they like, hence this new blog post.