Well-resolved images of galaxies (as seen at Galaxy Zoo) contain a lot of "morphological" information that is currently not being extracted and used. This is because fitting very flexible yet physically meangful models to high signal to noise ratio data of complex galaxies is hard! We'd like to advance astronomy by enabling a lot of new galaxy astrophysics parameters - such as the spiral arm winding tightness, star formation clumpiness, dust texture, and so on - to be inferred from our beautiful data.
GalaxyCraft is a web-based modeling tool that will enable anyone to choose a galaxy from the zoo, and make an artist's impression of it using a standard toolkit of simply parametrized but very realistic components: bulges, disks, spiral rms, dust, and so on. When the lovingly crafted model looks right, then the parameters of that model will already constitute a measurement of the new parameters. Collecting many such measurements from a small crowd of galaxy craftspeople should improve the accuracy and enable uncertainties to be estimated. A better approach would be to enable automated likelihood maximization and exploration initialized with a crafted model, via some sort of big red "optimize" button.
- Textured galaxy model suggested by Groeneboom & Dahle (2014)
- Galaxy ray-tracing developed at Shadertoy
- Javascript hacked at jsfiddle
- Otavio Good
- Phil Marshall
- Carl Gorringe
- Dave Johnson
Science Hack Day, San Francisco, 2014.