Originally developed by Jer Thorp and Mark Hansen for nytlabs, Cascade is a tool that affords the ability to see how a link is disseminated through retweets on Twitter. It presents this data as a tree structure allowing users to navigate the dissemination of a link over time by moving through parent-child relationships. The first version was written in Processing and was specifically designed to be installed in the lab. I was brought on the project to port it to something that could appear in a web page.
Using a similar strategy of what I employed in the creation of Delta, I made extensive usage of shaders to off-load a lot of the processing time devoted to moving points around. Each vertex contained all the information needed to describe its location in different contexts and modes so that despite the in-memory static geometry the points could animate and tween individually based on their specific vertex attributes. This amounted to an optimized version that could individually animated hundreds and thousands of tweets inside the browser with WebGL.