First, one pretty picture: we searched the web for "Comet Holmes" using Yahoo Image Search, resulting in over 1000 images of the sky from all over the web. We ran Astrometry.net on each one, and stacked all the images on the sky. From left to right, you can see the comet explode and become very bright and heavily imaged as it moves across the sky. The grid lines show RA,Dec. In a recent paper we show that we can fit a three-dimensional gravitational orbit to these images, showing that we can do science with this data set comprised of images of extreme heterogeneity and unknown provenance.
Postdoctoral Research Associate
Astrophysics Sciences
Princeton University
Room 007
Peyton Hall, 4 Ivy Lane
Princeton, NJ 08544-1001
Phone: 609-258-3809
My email address is username dstn, at the domain astro.princeton.edu .
I'm a postdoc working with the software group:
I work on data-reduction software for imaging from the Large Synoptic Survey Telescope (LSST), the new Hyper-Suprime Cam on the Subaru Telescope, and archival work on the Canada-France-Hawaii Telescope Legacy Survey (CFHT-LS) and Sloan Digital Sky Survey (SDSS).I also work with David Hogg (NYU) on a variety of projects involving statistical inference, machine learning, Monte Carlo methods, and forward- or generative modeling for astronomical data.
I also work with Julianne Dalcanton on the Panchromatic Hubble Andromeda Treasury project, a huge Hubble Space Telescope project to tile a quarter of the Andromeda galaxy with 6-band photometry, from the UV to the near-IR (F275W and F336W using WFC3/UVIS, F475W and F814W using ACS/WFC, and F110W (1100 nm), and F160W (1600 nm) using WFC3/IR).
I got my Ph.D in the Machine Learning group at the University of Toronto, Department of Computer Science in 2009. You can read my thesis if you've got time to spare.
If you're interested in Astrometry.net, you can hit that site to find out how to use Astrometry.net or read the Astrometry.net paper (roughly chapter 2 of my thesis).
You can find my astronomy papers listed here.