LISA mbhb-cats-and-samps 🐈‍⬛ & 🌽

Data release supporting:

• Stars or gas? Constraining the hardening processes of massive black-hole binaries with LISA. Alice Spadaro, Riccardo Buscicchio, David Izquerdo-Villalba, Davide Gerosa, Antoine Klein, Geraint Pratten. arXiv:2409.13011.

Credits

You are welcome to use this dataset in your research. We kindly ask you to cite the paper above. If you want to cite this data release specifically, the DOI is: .

Data

Data need to be downloaded from the github release page. The total size is ~1.2GB in total: ~1.1GB for the catalogs available, ~100MB for the posterior samples of TheCatalog inference. We provide 1000 simulated LISA realizations, in catalogs/LisaCatalogFrame_*.h5. Similarly, posterior samples for each event analyzed are in samples/source_*.h5.

Requirements

To read catalogs and samples you just need the pandas python package.

LISA Catalogs

Reading a catalog is as simple as

import pandas as pd
# Read the catalog
catalog_number = 1
cat = pd.read_hdf(f'catalogs/LisaCatalogFrame_{catalog_number}.h5', key='events')
# Inspect events in the catalog
print(cat.head())

If you want to open TheCatalog, as we refer to it in the publication, use catalog_number=713.

Units and conventions

Catalog columns are source parameters used for injection, f_cut,Gas fraction, SNR, Detection.

Units for source parameters, f_cut, Gas fraction, are specified in the dictionary below:

dimensionsdict = {
 'DimensionlessSpin1': 'dimensionless',
 'DimensionlessSpin2': 'dimensionless',
 'EclipticLongitude': 'radian',
 'Gas fraction': 'dimensionless',
 'InitialOrbitalPhase': 'radian',
 'LuminosityDistance': 'parsec',
 'Mass ratio': 'dimensionless',
 'MergerTimeOrInitialOrbitalFrequency': 'second',
 'Polarization': 'radian',
 'Redshift': 'dimensionless',
 'RedshiftedMass1': 'solar mass',
 'RedshiftedMass2': 'solar mass',
 'SNR': 'dimensionless',
 'cosInclination': 'dimensionless',
 'f_cut': 'Hz',
 'sinEclipticLatitude': 'dimensionless',
 }

for their definitions, please refer to the paper.

SNR is either a float or NaN, for the reason specified in the Detection string:

• Yes: Source is detectable
• LowSNR: Source is not detectable, given the SNR threshold
• LowMassRatio: Source has too low mass-ratio, for the waveform to be used confidently
• OutOfBand: the source is undetectable because its signal is outside the LISA sensitivity band

TheCatalog event posterior samples

Reading posterior samples from an event of TheCatalog is as simple as

import pandas as pd
event_number = 1
# Read the samples from the first event 
samples = pd.read_hdf(f'TheCatalog_samples/source_{event_number}.h5', key='samples')
# Inspect the samples
samples.head()

Units and conventions are similar to those of injections, just with a different parameterization.

{'DeltaMu': 'dimensionless',
 'DimensionlessSpin1': 'dimensionless',
 'DimensionlessSpin2': 'dimensionless',
 'EclipticLongitude': 'radian',
 'InitialOrbitalPhase': 'radian',
 'LuminosityDistance': 'parsec',
 'MergerTimeOrInitialOrbitalFrequency': 'second',
 'Polarization': 'radian',
 'RedshiftedChirpMass': 'solar mass',
 'cosInclination': 'dimensionless',
 'sinEclipticLatitude': 'dimensionless',
 'logL': 'dimensionless'}

In addition we store the (unnormalized) log-likelihood column logL.