When I think fast, I think being able to rendering hundreds of thousands of data points in real time, kind of like 'Google Maps' but for Time instead of Space, you should be able to fly over and analyze vast datasets.<p>As a historical note, since this is an area near and dear to my heart, I shipped an open source library several years ago circa 2007 that attempted to do this by applying the 3D graphics concept of a "mip map" to time series, with several representations of the data at various LOD automatically computed by sampling and filtering. I demoed it at Google I/O 2008 rendering 1 million data points in real time. Location here: <a href="https://code.google.com/p/gwt-chronoscope/" rel="nofollow">https://code.google.com/p/gwt-chronoscope/</a>, it's written in Java using GWT, and also the same code runs on the server for Java2D rasterization, and on Android. It has a JS API and Microformat API as well.<p>Here's a screencast of it in action, including things like animation, markers, synthetic datasets (technical analysis), styling, and history undo/redo. <a href="https://www.youtube.com/watch?v=RLYNHQVIeNg" rel="nofollow">https://www.youtube.com/watch?v=RLYNHQVIeNg</a><p>I eventually wanted to get to a point where people could collaborately analyze data and tell stories about it. This screencast shows "Timelord" which was a integrated with social networks and supported a concept called "Micro-presentations", using a markdown-like format to allow people to write animated two-way video/chart synchronized stories. In this video, Al Gore's movie "An Inconvenient Truth" is synchronized to a Chart of CO2 and Temperature.<p><a href="https://www.youtube.com/watch?v=s5Y21t0u_Zw" rel="nofollow">https://www.youtube.com/watch?v=s5Y21t0u_Zw</a>