The Nice Observatory on the French Riviera (Observatoire de la Cote d'Azur or OCA for short) uses Tao Presentations to display simulation results that contain about one billion data points.
For several years, Patrick Michel and his team have run simulations of asteroid collisions. These simulations, performed in collaboration with Derek C. Richardson from the University of Maryland (USA) run for several days, and generate huge data files containing the location of hundreds of thousands of particles over a period of time. A typical run will for example study 250,000 particles over 4,000 frames. That's one billion data points.
How do you understand what this data shows? Until now, the process was to plot individual frames and generate a movie. This process took several hours, if not days. And it resulted in a single viewpoint being explored. To see the data from another angle, you had to generate a new movie, which meant waiting a few hours or a few days again.
Using Tao Presentations, the OCA is now capable of viewing this data interactively, moving forwards and backwards in time as they wish, but also spinning around the particles, zooming, and so on. The video below shows what it looks like, for a simulation published recently. The simulation shows how the asteroid Itokawa may have been formed by gravitational reaccumulation of small fragments after the disruption of a large asteroid (Michel and Richardson 2013, Astronomy and Astrophysics 554, L1-L4) :
Because they were now able to circle around the data very quickly, researchers could now look at it from a large number of angles. And guess what? Seeing is understanding. Very quickly, new structures emerged from the data, such as the star-shaped pattern you can see in the video. Bringing interactivity into data visualization helped research progress faster.
This way of showing the data is only one of the many possible presentations. This particular one was created by Taodyne as a proposal for how to show and explain that research to the public, but the OCA also developed their own presentations for researchers. Since Tao Presentations is first and foremost a presentation platform, you can easily integrate big data visualization right into a larger and more general presentation.
To achieve their objective, the OCA created a C++ module that reads huge data files and plots the corresponding data using point clouds. To accelerate software development, Taodyne provided the OCA with the source code for the PointCloud module. The OCA then published their module on Gitorious, under the GPL licence.
So now you have all the elements you need to show gigantic collections of data in an interactive way, and integrate that directly into a presentation that also explains what you are showing, why it matters, or what you plan to do next.
What will you do next?
Christophe de Dinechin