What GWOSC offers

Open data

The Gravitational Wave Open Science Center (GWOSC, https://gwosc.org) gives open access to the data of the LIGO, Virgo and KAGRA observatories. It publishes the calibrated strain data of the LIGO, Virgo, GEO600 and KAGRA detectors, from the first observing run O1 to O4b (April 2024 – January 2025, with LIGO Hanford, LIGO Livingston and Virgo), sampled at 4 kHz and 16 kHz. With the strain come the data-quality timelines, auxiliary channels of the instruments and their environment, alternate strain versions, and seismic data. The Event Portal (https://gwosc.org/eventapi) gives every confident detection of the GWTC catalogs, with its parameters and data, and the community catalogs. Large data sets are served through the Open Science Data Federation (OSDF, https://gwosc.org/osdf), and the gwosc Python client (https://gwosc.readthedocs.io/) queries all of it from code.

Tutorials

  • Learning Path (https://gwosc.org/path): a gentle introduction to gravitational-wave data, in two tracks. For teachers and students: introductory videos, a waveform fitter for GW150914, sources and signals, an interactive signal-processing tutorial (https://gwfilter.streamlit.app) and the GW Data Quickview app (https://gw-quickview.streamlit.app). For researchers: the Event Portal, the data files and the analysis software.
  • Jupyter notebooks that run in the browser (Google Colab or mybinder) or locally: the Quickview notebook (https://github.com/gwosc-tutorial/quickview) to discover, read and plot the data; Introduction to GWOSC data (https://github.com/gwosc-tutorial/introduction_gwosc_data); and the code of “A guide to LIGO–Virgo detector noise and extraction of transient gravitational-wave signals” (https://github.com/gwosc-tutorial/Data_Guide).
  • Bulk data: tutorials to find and download strain files, by hand, from code, or for a whole observing run.
  • Community tutorials, such as the Gravitational-Wave Explorer (https://rachellanggin.com/tutorials/) beginner’s guide (waveforms, merger physics, parameter estimation).

All tutorials: https://gwosc.org/tutorials/

Gravitational Wave Open Data Workshops

Every year since 2018, GWOSC runs the Gravitational Wave Open Data Workshop, a free course in gravitational-wave data analysis: lectures by members of the LIGO, Virgo and KAGRA collaborations and hands-on Python sessions on the open data, from the basics of the signals and the noise to parameter estimation and population studies. It is followed online and in local hubs around the world, and the lectures and notebooks remain available afterwards.

  • April 20–23, 2026 — course: https://gw-odw.thinkific.com/courses/odw2026
  • May 12–14, 2025 — course: https://gw-odw.thinkific.com/courses/odw2025
  • All workshops and their material: https://gwosc.org/workshops/

Software

GWOSC lists the open-source software of the gravitational-wave community, most of it installable from the IGWN Conda distribution (https://gwosc.org/software/):

  • Data access: the gwosc client, GWpy, the Frame Library, and examples for MATLAB and C.
  • Searches: PyCBC, GstLAL, coherent WaveBurst (cWB), Omicron.
  • Parameter estimation: bilby, BayesWave, RIFT, LALSuite.
  • Sky localization and alerts: ligo.skymap (with BAYESTAR), PyGCN.
  • Populations: Bayesian parametric population models of the merger rate.

 

What GWOSC+ offers

GWOSC+ is the virtual access installation of the European Gravitational Observatory (EGO) in the ACME project. It gives free remote access to the GWOSC data and to gwtc_analysis, a tool to analyse them.

gwtc_analysis

gwtc_analysis is an open-source Python package that turns the published Gravitational-Wave Transient Catalogs into science, from single events to the whole population. It analyses every public catalog, from GWTC-1 to GWTC-5.0 and the GWTC-4.1 update (observing runs O1 to O4b), and keeps serving the catalogs as they are released: GWTC-6, from the O4c observing run, is coming soon. It is designed for newcomers and multi-messenger astronomers as well as experienced gravitational-wave analysts.

It works directly on the official data products. Parameter-estimation samples, skymaps, strain data and search-sensitivity injections are retrieved automatically from GWOSC and Zenodo, with the release versions resolved at run time. A single command goes from the public data to a complete report, with figures and tables, that lists every data file it used and the options of the run.

Besides Zenodo and the other public sources, the files can be read from two mirrors: an S3 bucket and a public usegalaxy.org history. A missing file or a failed transfer falls back on another source, and network errors are retried.

gwtc_analysis reproduces published results of the LVK Collaboration:

  • the Hubble constant from the binary-black-hole mass spectrum (spectral siren) and with a galaxy catalog (dark siren), matching the GWTC-4.0 cosmology paper;
  • the Hubble constant from GW170817 and its host galaxy NGC 4993, alone and combined with the dark siren;
  • Hawking’s black-hole area theorem tested with GW250114;
  • the GWTC-4.0 search for strongly lensed pairs in O4a;
  • the coincidence of GW170817 and GRB 170817A, and the bound it sets on the speed of gravity;
  • the gravitational-wave background of unresolved compact binaries, in agreement with the LVK prediction and compared with the latest upper limits of the stochastic searches.

Analysis modes

gwtc_analysis is run as gwtc_analysis MODE [options]. It has 18 modes.

  • Exploring the catalogs
    • catalog_statistics: detector-network participation, distributions of the source parameters, sky-localization areas, and the energy radiated and remnant spins.
    • event_selection: events selected by mass, distance and spin, or by class of sources: neutron-star binaries, the lower mass gap, and hierarchical-merger candidates in the pair-instability gap.
    • search_skymaps: the events whose sky localization contains a given position; with –counterparts, the gamma-ray bursts (Fermi-GBM, Swift) and IceCube neutrinos coincident with each merger.
    • check_catalogs: whether GWOSC and Zenodo have published a catalog that the installed version does not cover yet.
    • zenodo_releases: the Zenodo release versions of each catalog, to pin an analysis to one of them.
  • Single events:
    • parameters_estimation: everything about one event: its posteriors, the strain of each detector against the best-fit waveform, time-frequency maps, matched-filter signal-to-noise ratios, the evidence for higher multipoles and orbital precession, the remnant and its energetics, the tidal deformability of neutron stars, and the 3D sky map with candidate host galaxies.
    • counterpart: where an electromagnetic counterpart sits in the gravitational-wave posterior: its sky position, its distance against the host redshift, and the viewing angle.
    • build_unofficial_pe: a parameter-estimation bundle of GW170817 built from its public releases, for the modes that need one.
  • Populations and cosmology:
    • rates: binary neutron star, neutron star–black hole and binary black hole merger rates, corrected for selection effects with the LVK search-sensitivity injections, together with the redshift evolution of the merger rate.
    • hubble_constant: the Hubble constant by four methods (–method): from the binary-black-hole mass spectrum (spectral siren), from the mass spectrum and a galaxy catalog (dark siren), from an event with an identified host (bright siren: GW170817 and NGC 4993, or the candidate flare of GW190521), and the joint posterior of independent results.
    • galaxy_catalog: the line-of-sight galaxy catalog of the dark siren (GLADE+ by default) for icarogw, on a laptop or as Slurm jobs on a cluster.
    • spin_population: the effective-spin population of binary black holes and its correlation with the mass ratio.
    • stochastic: the gravitational-wave background of unresolved compact binaries, against the upper limits of the LVK stochastic searches.
    • lens_pairs: every pair of binary-black-hole events ranked for strong lensing by the overlap of their posteriors and their time delay.
    • Note: The rates and hubble_constant modes take the LVK population models with –mass-model: plp (Power Law + Peak), mltp (Multi Peak) and fullpop4 (FullPop-4.0, which describes all compact binaries at once).
  • Fundamental physics and neutron stars:
    • area_law: Hawking’s black-hole area theorem tested with GW250114, inspiral against ringdown.
    • neutron_star_eos: the neutron-star equation of state from GW170817 and GW190425 jointly (tidal deformability and radius of a 1.4 solar-mass star).
    • scalar_cloud: whether a binary black hole is better fitted with an ultralight scalar-field cloud around it (GW190728).
    • continuous_waves: the LVK continuous-wave upper limits (known pulsars, all-sky O4a, O3 boson clouds) and the boson mass inferred from GW190728.

 

Figure: Illustration of example results

Open and reproducible

gwtc_analysis is developed openly on GitHub, tested continuously, and documented with a user guide, the methods and their references at https://danielsentenac.github.io/gwtc_analysis/. It installs from PyPI (https://pypi.org/project/gwtc_analysis/) or conda-forge (https://anaconda.org/conda-forge/gwtc_analysis), runs anywhere with the gwtc-tool Docker image (https://hub.docker.com/r/danielsentenac/gwtc-tool), scales to computing clusters with Slurm, and every release is archived on Zenodo (DOI: https://doi.org/10.5281/zenodo.23077807).

It builds on the community’s standard libraries: gwosc, GWpy, PESummary, ligo.skymap, PyCBC, bilby and icarogw.

Links

Link to GWOSC/GWOSC+ website