Open-source astrostatistics software projects
The following projects are hosted at GitHub:
- The batse5bp package provides persistant local access to BATSE GRB data hosted at the CGRO SSC. It accesses both the BATSE GRB catalogs, and a significant amount of the detailed GRB data (e.g., time series and spectral data). Sphinx documentation
- The inference package is a collection of Python modules implementing a variety of methods targeting the statistical inference problemsand the statistical modeling styleof the physical sciences. Documentation forthcoming. Please pardon our dust! Inference is being significantly overhauled as part of updating it from Python 2.7 to Python 3. The GitHub-hosted version includes some work-in-progress; the setup.py script installs only the parts passing basic compatibility tests.
- A more "Pythonic" interface for the Stan Bayesian probabilistic modeling language, via PyStan. Initially developed for the XXVI Canary Islands Winter School of Astrophysics, and subsequently extended for use in Cornell University's STSCI 4780 — Bayesian Data Analysis: Principles and Practice. Public documentation forthcoming.
- A C++ framework for GPU-accelerated hierarchical Bayesian inference with two-level multilevel levels, using NVIDIA's CUDA library. The motivating applications are cosmic demographics, such as modeling number-size ("log(N)-log(S)") distributions, luminosity functions, initial mass functions, etc., accounting for measurement error and selection effects. A paper describing the methodology and software is in preparation for Astronomy & Computing.