nithyanandan / 11yr_stochastic_analysis

Repository containing models and running scripts corresponding to the 11-year Stochastic Background analysis paper

11yr Stochastic Analysis

This repo provides data and software to reproduce results from the NANOGrav's 11 year stochastic background analysis paper (see paper, dedicated webpage). The data is a subset of the full 11 year data release available at data.nanograv.org. NANOGrav primarily uses Python for data analysis, and this repo makes use of a Jupyter notebook.

For the paper most of the analyses were conducted using PAL2 and NX01. These packages have since been superceded by enterprise, which is used here.

The bayesephem directory contains a Jupyter notebook that reproduces all computations necessary to set up the BayesEphem model of solar-system-ephemeris uncertainties, which is used to make this analysis robust to ephemeris errors. BayesEphem is discussed in the NANOGrav paper "Ephemeris uncertainties vs. pulsar timing arrays".

There are two ways perform the analysis: using Docker and using a local installation of software.

using Docker

The docker image provides the data and a minimal installation of all of the software you need for the analysis. To run the analysis simply:

• Pull the latest docker image.

docker pull nanograv/nano11y-gwb

• Start a new container. Make sure to change the path to a local directory for you to store your results (following the -v option). This will ensure that any data products are persistent, even if you delete the docker container. You can change the display port if you are already using 8888 for something else.

docker run -it -p 8888:8888 \
 -v /path/to/local/dir/:/home/nanograv/local_data/ \
 -u nanograv nanograv/nano11y-gwb

the default action for docker run will launch jupyter in the container home directory.

• Copy and paste the output URL into your browser, and open the analysis.ipynb notebook

using a local install

To run the analysis you will need to have enterprise and PTMCMCSampler (and their dependencies) installed. enterprise requires either libstempo or PINT to read in the data from .par and .tim files. libstempo in turn requires temop2, a sometimes unruly pulsar timing package.

With all of that ready you can clone this repo, and begin.

git clone https://github.com/nanograv/11yr_stochastic_analysis.git
cd 11yr_stochastic_analysis
jupyter notebook