I'm Léonard Seydoux, a geophysicist and data scientist at the institut de physique du globe de Paris as a junior professor at Université Paris Cité.
I mostly study earthquakes and volcanic eruptions. My research analyzes subtle geophysical signals, including slow earthquakes, using artificial intelligence and machine learning. I combine high-performance computing, array-based seismic analysis, and advanced signal processing to link these signals to major geological events. I focus on AI-driven hazard assessment, reconstruction of sparse or historical datasets, and multiscale monitoring of fault and volcanic systems, while integrating software development, teaching, and collaborative supervision to connect computational geophysics with practical understanding of Earth's dynamics.
There are several open-source projects that I develop and maintain. If you find them useful, please consider giving feedback, reporting issues, or contributing to their development!
| Project | Description |
|---|---|
 |
Covseisnet is a Python package for array signal processing, with a focus on data from seismic networks. The core mathematical object of the package is the network covariance matrix, used for signal detection, source separation, localisation, and plane-wave beamforming. The signal detection and processing methods are based on the analysis of the covariance matrix spectrum. The covariance matrix can be used as input for classical array processing tools such as beamforming and inter-station cross-correlations. |
 |
Scatseisnet is a Python package for seismic signal detection and classification using deep learning. It is built on top of Covseisnet and leverages covariance matrix analysis for efficient seismic data processing. The package includes pre-trained models for various seismic signal types, including earthquakes, volcanic tremors, and anthropogenic noise. Scatseisnet is designed to be user-friendly and easily integrable into existing seismic data analysis workflows. |
 |
PyCPT-city is a Python package for bringing color palette tables (CPT) from the cpt-city to Matplotlib. It allows users to easily apply a wide range of color palettes to their visualizations, enhancing the aesthetic appeal and interpretability of geospatial and scientific data plots. The package supports various CPT formats and provides functions for loading, converting, and applying these color palettes in Matplotlib. |
 |
PyGMRT is a Python package for automating the fetch data from the Global Multi-Resolution Topography (GMRT) Synthesis database. The package is designed to facilitate the integration of GMRT data into geospatial analysis workflows, making it easier for researchers and scientists to access and utilize this valuable resource for their studies. |
| Contributed | Description |
 |
Beampower is a package for beamforming (or backprojection) of seismic signal for event detection and location. The Python wrapper can call the C (CPU) or CUDA-C (GPU) implementation. |