Posted by Daniel Mashburn on August 22, 2014
Each weekly meeting covered roughly a single chapter from Statistics, Data Mining, and Machine Learning in Astronomy by Ivezic, Connolly, VanderPlas & Gray, all organised by Laleh Sadeghian (postdoc). Primarily graduate students and post-docs in CGCA have taught the course, but attendance has included every level of learner from undergraduate to professorial. Learning from my peers was a great opportunity, as no one is a true expert; silly questions were never ignored. When I had a turn to present, I also found the experience more rewarding. Because my audience had a similar mastery of the subject material, except of course for my particular chapte there was no need to fear that you were talking above someone’s head or beneath another’s expertise.
Many of the things I learned in class, I have put into use readily. The chapter on bootstrapping, presented by Megan DeCeaser, takes a single measurement from a few points, and performs the calculation an extra thousand times by randomly resampling your data set. Or, if you I have an over-abundance of data giving me grief, then simply reduce those discreet points to a simple hierarchal structure with an ‘extreme deconvolution,’ (see figure 1).
The advancement of computing technologies in recent decades, and its continued acceleration, ensures a need for well-trained programmers in the field of astronomy. Taking a summer course in the CGCA is a relaxed yet intensive way to learn, teach, and meet some new people who share a common goal in education and exploration. If you think hosting a summer study course is right for you, or maybe you just want to keep your research skills sharp in the long and hazy summer evenings, then get ready for next year!
Figure 1: An example of extreme deconvolution of stellar data. Image taken from http://www.astroml.org/book_figures/chapter6/index.html