Here at Rdio, we have a hack day event once a quarter. During this event, which lasted two days, we work on anything we want related to Rdio. I chose to implemented a fun visualization algorithm. I wrote a blog post about it on the official (and new) Rdio engineering blog.
For Rdio's quarterly hack day, I decided to focus on the visualization of similarity between people. It's easy to compare two people and show how similar they are. You can show a similarity score, you can show the actual artists their collections have in common, or you could even have some fun graphics that represent similarity such as a traffic light. But what about three people? ten people? a hundred?
With multiple people, the most obvious choice is place them on a 2D plane, with similar people being close to each other and dissimilar people being far apart.
Read the rest on Rdio!