DATA VISUALIZATION AS THE DRIVER OF SCIENTIFIC DISCOVERY
Visualization has traditionally been the last step in the scientific process, almost an afterthought. Visualization, however, can take
advantage of the unparalleled ability of the human brain to identify patterns in complex datasets. Here you will see some of my efforts
in using data visualizations as a driver of scientific discovery. My goal is to make scientific visualizations that are not only pretty but also informative.
Fast self-adaptive volume rendering
This post shows some experiments I have been doing with pseudo volume rendering using Delaunay tessellations.
While there are lots of techniques that can be used to display large point datasets there are very few techniques
for visualizing sparsely sampled points. This is a very common situation when dealing with (astronomical) real
data like galaxy redshift catalogues.
I have been working for quite some time in a very cool 3D tiling engine called DeepSky. It is a similar
idea to google maps but here you stream 3D “tiles” as XML files with positions and other information of
galaxies inside the tile. Next the galaxy images are streamed from the SDSS cutout service and cached to
disk for faster future retrieval.
DeepSky is fully functional although no user-friendly as some options are written directly in the source code.
It was written in processing but eventually
I will port it to c++ using openframeworks.
This is a fly-through animation I made in collaboration with Mark SubbaRao (Adler Planetarium). The movie has been
featured in several documentaries and went viral in youtube when it was posted as part of the SDSS DR9 release.
Some time ago I started collaborating with Nick Scoville, P.I. of the COSMOS survey on visualizations of
the COSMOS survey. We had worked mainly in visualization of the dataset consisting of galaxy images and photometric
redshifts. Here is one movie we made using approximately a quarter of a million galaxy images. I used a galaxy
template like the one used in the SDSS fly-through.