(Jeffrey D. Scargle, Tamas Budavari)
This poster provides an overview of open-source software packages addressing two challenging classes of astrostatistics problems arising in high-energy astrophysics. (1) Time Series Explorer (TSE) is a collection of software in Python and MATLAB for exploratory analysis and statistical modeling of astronomical time series. It comprises a library of stand-alone functions and classes, as well as an environment for interactive exploration of times series data. We summarize key capabilities of this emerging project, including new algorithms for analysis of irregularly-sampled time series. (2) CUDAHM is a C++ framework for hierarchical Bayesian modeling of cosmic populations, leveraging graphics processing units (GPUs) to enable applying this computationally challenging paradigm to large datasets. CUDAHM is motivated by measurement error problems in astronomy, where density estimation and linear and nonlinear regression must be addressed for populations of thousands to millions of objects whose features are measured with possibly complex uncertainties, potentially including selection effects. An example calculation demonstrates accurate GPU-accelerated luminosity function estimation for simulated populations of one million objects in about two hours using a single NVIDIA Tesla K40c GPU.