Rubinium.org

Google has added a feature to YouTube for navigation of related videos. Video bubbles are placed by attraction. If it works like our algorithm at work for displaying networks and flow graphs here’s the general idea behind the algorithm. Videos with a higher relation ranking have bubbles which are assigned a higher attraction. There is a certain amount of built-in repulsion to space all bubbles out. Then position is determined by summing all positions of other bubbles weighted by the attraction and repulsion values. Ok, so it is a bit more complicated than that when you consider all the optimizations weighting functions but thats the starting point.