AGU 2021 • 17 December 2021
Feel free to screenshot/share/reuse/remix this presentation
We solve inverse problems in Geophysics
and build open-source software
Moderate compute and data
Example: 3D density distribution from gravity disturbances
Build as much as possible into open-source from the start
New methods won't be used without an implementation*
* Doesn't even have to be a good one
Paper == GitHub repo
Code + Data + Figures + Text
Private at first, public on submission, archived on acceptance
Could be better:
Notebook review on GitHub 😔
Notebooks + Make
Barrier to entry
- Matlab
- Fortran/C++ on HPC
- Python on laptop (Jupyter)
- Python on the cloud (Pangeo)
- Bash (csh 🧙) scripts
- Google Earth Engine
- Commercial GUIs (ArcGIS, Petrel)
- OSS GUIs (QGIS, OpendTect)
A lot of effort to replace existing stack
Interoperability between new and established tools
Education of scientists on new ways of working
Trend towards Python + GitHub + Jupyter
Code + OSS ➡️ Openness + Reproducibility ❤️
Find out more:
Source code for this presentation:
github.com/compgeolab/agu2021
Unless otherwise noted,
the contents of this presentation are
licensed under the
Creative Commons Attribution 4.0 International License