1. 10% of podcasts in our database have social accounts (e.g., Twitter) & relevant urls (e.g., Spotify, YouTube...). More to come. This sounds trivial, but I spent some time building internal tools to make it easy to collect these information.
2. Improved ranking algorithms for the search engine, which involved some data cleaning & small tweaks on weights of different ranking signals.
3. A cleaner version of the API marketing page.
4. Listen API returns more data now, including episode-specific images, podcast social accounts & urls, classified ads, and listennotes_url.
5. Made a Paw file for API documentation.
6. Work in progress: a simple web app for Listen API demo, which will be open sourced soon.
7. Work in progress: some Listen Later improvements, which will be my main focus in October.