This gist contains some code used in Meszaros et al. (2015) pertaining to the analysis of multiple populations in globular clusters using the extreme deconvolution algorithm. The necessary data files are available upon request.
The basic use of the code is to run:
IDL> analyze_gc, gc=2, ngauss=2, outfilename='m2pho_2comp.fits'
This runs the analysis on a given globular cluster (see read_gc.pro for the mapping between gc
and globular cluster). The produced output file is a fits file that has the ID, the component that the XD code assigns (xdpop), the probability of that assignment (logp, actually the natural log of the probability), and the population that was in the original data file from a preliminary analysis of membership probability. Information on the best-fit amplitudes, means, and covariances of the populations is printed directly.