Gravitational Wave Open Data Workshop 2026

A complete hands-on journey through gravitational-wave data analysis using open LIGO/Virgo data.

From raw detector strain to event detection, parameter estimation, and continuous-wave searches.

Overview

This repository contains completed solutions, analyses, and documentation for the Gravitational Wave Open Data Workshop (GW-ODW) 2026.

The workshop covers the full gravitational-wave data analysis pipeline:

• Accessing public detector data from GWOSC
• Generating theoretical waveforms
• Signal processing and noise whitening
• Matched filtering and significance estimation
• Bayesian parameter estimation using Bilby
• Population inference and continuous-wave searches

In addition to the official tutorial notebooks, this repository includes detailed solutions to all four data challenges, along with professional markdown summaries, figures, and interpretations.

Highlights

• Completed 5 core tutorial modules
• Solved 4 progressively advancing data challenges
• Performed Bayesian parameter estimation with Bilby
• Generated and interpreted Q-transforms, SNR time series, and posterior predictive checks
• Detected merger events and glitches from continous data
• Explored advanced topics including:
  • LALInference
  • Population inference
  • Frequency-Hough continuous-wave searches

Run the notebooks interactively in Google Colab

Data Challenge

The data challenge synthesizes the concepts introduced in the tutorials and provides end-to-end analysis exercises.

Challenge	Focus Area	Key Outcome	Guide
1	Time-frequency analysis	Identified a clear gravitational-wave chirp
2	Matched filtering	Recovered a weak signal and measured SNR
3	Multi-detector coincidence	Confirmed the event independently in H1 and L1
4	Event search and parameter estimation	Detected multiple events, classified glitches, and inferred source masses

Challenge 1 — Chirp Detection

A Q-transform spectrogram was used to identify the characteristic chirp signature of a compact binary merger.

Challenge 2 — Matched Filter SNR

Matched filtering recovered a weak signal buried in detector noise and produced a clear SNR peak at the merger time.

Challenge 3 — Multi-Detector Coincidence

Matched filtering was performed independently on Hanford (H1) and Livingston (L1) data. Consistent SNR peaks at the same coalescence time confirmed the astrophysical origin of the signal.

Challenge 4 — Event Search, Glitch Detection, and Parameter Estimation

Candidate Event Detection

A long-duration dataset was scanned to identify candidate events. The earliest confirmed detection near 826.43 s was selected for Bayesian parameter estimation.

Glitch Detection

Representative Instrumental Glitch — Koi Fish

As an additional analysis, representative detector artifacts were visually compared with the Gravity Spy taxonomy and assigned likely morphological classifications.

Example classifications:

• Koi Fish
• Blip
• Tomte
• Fast Scattering
• Whistle
• Low Frequency Blip
• Paired Doves

Parameter Estimation Result

For the earliest confirmed event, Bayesian inference yielded component masses with a 90% credible interval of

$$ 28.70 \le m \le 28.76,M_\odot $$

consistent with an equal-mass binary black hole merger.

Tutorials

The workshop tutorials provide the theoretical and computational foundation for the challenge analyses.

Tutorial	Topic	Guide
1	Accessing Open Data
2	Generating Waveforms
3	Signal Processing
4	Searches
5	Parameter Estimation

Example Outputs

Binary Black Hole Waveform

Q-Transform of a Gravitational-Wave Event

Posterior Predictive Check

Extension Topics

Optional advanced notebooks covering research-grade methods:

• Parameter estimation with LALInference
• Hierarchical population inference
• Continuous-wave searches using the Frequency-Hough transform

Frequency-Hough Map

Repository Structure

tutorials/        # Core workshop tutorials
data_challenge/   # Completed challenge solutions
assets/           # Shared images and graphics

Running the Notebooks

The easiest way to explore the notebooks is through Google Colab.

Click the Open in Colab badge at the top of this README to launch the repository in a hosted environment with no local setup required.

Local Setup (Optional)

pip install -r env/requirements.txt

or

conda env create -f env/environment.yml
conda activate gw-odw

Technologies Used

• Python
• NumPy
• SciPy
• Matplotlib
• GWpy
• PyCBC
• Bilby
• PESummary
• LALSuite
• PyHough

Official Workshop Repository

The original workshop materials and course content are available at:

• 🌐 Workshop Portal: GW Open Data Workshop 2026
• 💻 Official Repository: gw-odw/odw: Materials from GW Open Data Workshop

Acknowledgements

This work is based on the Gravitational Wave Open Data Workshop organized by the LIGO–Virgo–KAGRA Collaboration and the Gravitational Wave Open Science Center (GWOSC).