Data mining
Last updated on 16/05/2026
The IT dedicated to the intensive exploitation of large volumes of data is mainly related to the research activities of the Ondes team, activities related to methods based on ambient noise. This expertise is at the interface between geophysicists, HPC specialists and databases in the context of national and European projects.
The codes and associated documentation developed as part of the activity "Data Mining" are accessible in the Forge GitLab GRICAD (or the Forge OSUG for projects not yet migrated)). It should be noted that the software forges and/or documentation of some of these projects are restricted access.
WIN->MSEED format conversion tools / pre-processing / correlations / doublets & inversion (Whisper and F-Image projects) :
Development and deployment of a code for calculating seismic phase velocity variations (dv/v), implemented on Japanese data from the Hi-Net network, data from the YH network (Dense Array for North Anatolia), SCEDC Californian data, ...
- WIN to MSEED conversion tool code
- wiki and journal of Whisper project , Japanese NIED data documentation, data preprocessing documentation and code
- correlations / doublets / inversion tools documentation and codes
Associated publications :
- doi:10.1038/s41467-022-32194-3
- doi:10.1002/2017JB014307
- doi:10.1029/2019JB017803
- doi:10.5194/egusphere-egu2020-15195
- doi:10.1016/j.epsl.2021.116894
Beamforming tools (Imag’In, RESOLVE projects) :
This interdisciplinary project (IGE, ISTerre, LIG) studies data acquired on the Argentière glacier in spring 2018 ( RESIF ZO_2018 dataset ). The objectives are to develop a code for calculating MFP (Matched Field Processing) locations, to perform this calculation on the CiGri computing grid, and to develop a tool for visualising the multiparametric results (x, y, z, c, f, t) obtained from these locations. The tools developed as part of this project are used for the SPIN ITN project (Cuolm da Vi landslide, ETH Zürich), the Tête Rousse Glacier project (ISTerre Fault Mechanics team), the SEIS-ADELICE project (Astrolabe Glacier-Antarctica, IGE) and the SLIDE, REASSESS and SEISMORIVGLA projects (Isuunguata Glacier, Greenland, IGE). The processing of data from the Isuunguata Glacier in Greenland served as a basis for testing the large-scale data processing campaign via the CiGri grid on the two new (2025) supercomputers kraken-cpu and kraken-gpu ( freeride MFP-CiGri in summer 2025 ).
- wiki RESOLVE
- documentation and code
- documentation MFP code
- documentation and code (old version, no longer maintained)
Associated publications :
- doi:10.1073/pnas.2023757118
- doi:10.1093/gji/ggaa411
- doi:10.1029/2021GL095996
- doi:10.1029/2023JF007280
- doi:10.1785/0220200280
- doi:10.1093/gji/ggv069
- doi:10.1093/gji/ggag103
- doi:10.5194/egusphere-egu26-10595
- Dense Seismic Array Monitoring of the Glacier Tongue of Isunnguata Sermia, West Greenland, Paris et al., submitted to Seismica (27 Jan 2026)
Visualization tools for beamforming outputs (collaboration with R. Blanch and M. Ortega from LIG)
Optical Correlation project
The objective of this project is to prototype a deterministic and analytical method for measuring sub-pixel scale displacements between two 2D images (e.g. measuring ground displacements with an amplitude smaller than a pixel). Through successive iterations on pairs of local sub-images, a discrete 2D correlation is calculated and fitted using a continuous function : an objective function dependent on various parameters (horizontal and vertical displacements, anisotropy, amplification factor, etc.). The best fit provides information on the optimal parameters selected to describe the ‘displacement’, as well as the uncertainty associated with the optimisation. The test data sets used are of different types : PIV (Particle Image Velocimetry : data from the ImVort project, LEGI wave channel), GeoTiff images of the 2013 Balochistan earthquake (Pakistan).
- documentation and code and wiki
- Example of videos of displacement maps in the context of the 2013 Balochistan earthquake
seedpsd project (collaboration with ISTerre-Faults team and ISTerre-Waves team)
data container evaluation project (collaboration with ISTerre-Waves team, EPOS-Fr, Earthscope, GFZ)
- [https://gricad-gitlab.univ-grenoble-alpes.fr/tests-isterre/data_container]
- [https://gricad-gitlab.univ-grenoble-alpes.fr/tests-isterre/data_container/-/tree/hdf5-to-zarr]
Associated publications :
This project is related to the Prototyping tools to link RESIF data center data and CIMENT-GRICAD HPC infrastructures (Resif-Summer-Ciment project and associated code)
PYCORR project (collaboration P. Boué et L. Stehly from ISTerre-Waves team)
Template Matching Tools (EventDetection project)
Associated publication :
NoiseCorr_DBF : Correlation tools and double beamforming (sanjacinto project) :
Double-beamforming and correlation calculations using seismic acquisition data around the San Jacinto Fault (Southern California). DOI ZG_2014 dataset
Associated publication :
Time error detection tools on dense networks (IWORMS project) :
The objectives of this interdisciplinary project (LJK, ISTerre, GriCAD) are (1) to implement a new method for correcting dating based on the intercorrelation of ambient seismic noise (2) to apply this processing to continuous data streams.
Associated publication :
Tools for the manipulation/reorganization of datasets of valued data in HDF5 format (Utils project) :
Conversion format FCNT to MSEED tools (IMPROVE project : Krafla et Harmalière) :
Tools for performing flow velocity and particle concentration measurements based on Acoustic Particle Image Velocimetry (ImVort project = Imagerie-Vorticité) :
Others :
Link to internal HDF5 training materials for RESIF-EPOS-Fr, ISTerre-SIG service, and IPGP staff
Link to grid computing CiGri training and presentation materials
Link to the training courses : tools for data processing, software development and computing (updated as sessions progress, full version available on request)
The business expertise of the technical staff involved are :
– optimization of sequential codes (numerical methods, choice of languages,...)
– application parallelization (MPI, OpenMP)
– application deployment
– grid calculation (CiGri v3)
– iRODS : transfer techniques, metadata management,...
– IO parallel
– HDF5 data format
– signal processing : hole management, data decimation,...
– Fortran / C / Python3 / Shell Bash
Contacts for the activity ’Data mining’ :
– Michel Campillo, Philippe Roux, Florent Brenguier : researchers, F-Image project manager, RESOLVE, Pacific
– Albanne Lecointre, IR CNRS BapE, team ISTerre-Waves
Specific skills
– Fortran, C, Python3
– MPI, OpenMP, Grid calculation....
– Scientific computation libraries BLAS Lapack, Scipy, IntelMKL
– h5py, opspy, numPy, scipy, matplotlib, ...
– File format HDF5, SEED, (NetCDF3)
– data scan : msi
– format conversion : sac2mseed, win2sac_32
– metadata mining : rdseed
– Data type : seismological
Hardware and software resources backed by the activity’Data mining’
– Link to the CIMENT/GriCAD computing center
– Link to the ISTerre calculation means
Links with other IT activities and resources in ISterre and OSUG
==> ISterre Data Centre
==> Laboratory IT resources
==> OSUG Storage Center
Links with other ISTerre technical platforms
The federation
Intranet
