Short GRB Delay Times

Delay time distribution of short GRBs

This dataset contains all data and code used in the analysis of Zevin et al. 2022. Upon downloading the tarball 'short_GRB_delay_times.tar.gz', one can unpack it using the command 'tar -xzvf short_GRB_delay_times.tar.gz'. Below we describe all components of this dataset:

• 'bright_grbs_clean.ipynb': a jupyter notebook that performs the entire analysis of Zevin et al. 2022, and generates all numbers/figures from this work. 
• 'GRB_pcc.txt': a text file that contains the Probability of Chance Coincidence classification (gold, silver, bronze, inconclusive) for each GRB in the dataset. 
• 'samples/': this directory contains the Prospector posterior samples for each GRB host and was downloaded from the BRIGHT website. 
• 'figures/': directory that is populated by bright_grbs_clean.ipynb and contains figures used in the paper
• 'alpha_tmin_tmax.../': directories that contain the Dynesty nested sampling output for various runs that used different prior settings and different sample criteria. The files names 'label_samples.dat' contain the posterior samples. The four main runs used in this work are 'alpha_tmin_tmaxextended_GRB170817', 'alpha_tmin_tmaxextended_gold_GRB170817', 'alpha_tmin_tmaxextended_specz_GRB170817', and 'alpha_tmin_tmaxextended_gold_specz_GRB170817'. The run 'alpha_tmin_tmax_GRB170817' uses a truncated prior on the maximum delay time of 14 Gyr rather than 100 Gyr. 'alpha_tmin_tmaxextended' does not include GRB170817 in the analysis. Runs that end with '_only' perform the analysis with only the GRBs that are listed in the directory name. 
• 'interpolants/': directory that contains the precomputed interpolants of the DTD likelihood surface, which are used in our hierarchical inference. These grids can be recomputed in the Jupyter notebook, though are somewhat expensive to run (~5 hours). 
• 'merger_rate_evolution/': directory that contains a pickle file of precomputed star formation rate densities that are used to create Figure 3; these can also be recomputed using the Jupyter notebook.