GBT-dark-matter

Axion dark matter search using Breakthrough Listen GBT data

Original paper by Aya Keller et. al introducing the asymmetry analysis technique to search for dark matter using the Breakthrough Listen nearby star sample, and demonstrating the technique on a 100 MHz region of the L-Band

We are expanding this analysis to the complete L-Band, and eventually the complete dataset of L, S, C, and X-Band.

To adapt and/or run this analysis: git clone https://github.com/nicolewolff/GBT-dark-matter.git

Contact me at [email protected] for information about obtaining the datasets used for this analysis, or search for individual files here to use to test a specific part of this analysis.

To convert h5/filterbank data products to NumPy .npy files (in Python):

import blimpy as Waterfall
import numpy as np

wf = Waterfall()
data = wf.grab_data()
data = np.mean(data, axis=0)  # Average over the time axis
np.save('[filename].npy', data)

Running the analysis

Open the file config.ini. Edit the directory paths accordingly. Edit the settings for signal injection: start_frequencies represents the first frequency at which a signal is injected, and this injection is repeated every 50 MHz. (To change the signal modulation, edit inject_spaced_arg.py.) Edit the operations to control which scripts are run. For the full analysis, first, set the signal size to 0, and set only inject, uninjected_preprocess, uninjected_normalize, inject_template, preprocess_template, normalize_template to True. Then, set only inject, preprocess, normalize, and asymmetry to True.

To understand specific parts of the analysis, read the source code comments for each function.