[Incubator] TorchGeo interested in joining OSGeo

[Adam Stewart]

Hello,

My name is Adam Stewart. I'm a geospatial ML researcher <https://scholar.google.com/citations?user=IQ19q4AAAAAJ> and passionate open-source developer <https://github.com/adamjstewart>. I earned my B.S. in Science of Earth Systems at Cornell and my Ph.D. in Computer Science at UIUC, and currently work as a postdoc at TUM with a focus on machine learning for Earth observation. I created TorchGeo <https://github.com/microsoft/torchgeo>, and also help maintain the R-tree <https://github.com/Toblerity/rtree> library and Spack <https://github.com/spack/spack> package manager. I've contributed to most geospatial libraries, including GDAL <https://github.com/OSGeo/gdal/issues?q=author:adamjstewart>, rasterio <https://github.com/rasterio/rasterio/issues?q=author:adamjstewart>, and R-tree <https://github.com/Toblerity/rtree/issues?q=author:adamjstewart>, and maintain all geospatial build recipes in the Spack package manager.

I'm reaching out to you today to express interest in having TorchGeo <https://github.com/microsoft/torchgeo> join the OSGeo Foundation. TorchGeo is a PyTorch domain library for machine learning with geospatial data, especially remote sensing data. TorchGeo started as an internship project at the Microsoft AI for Good Research Lab and developed into my thesis dissertation with an active community of contributors from around the world. It builds on top of GDAL (via rasterio and fiona) and PyTorch to make it easy to work with any kind of geospatial data in a machine learning workflow. Data can be automatically reprojected and resampled to a common CRS and resolution, making it easy for machine learning researchers without geospatial skills to work with satellite imagery. 

Some statistics:

Stars: 2K
Forks: 240
Contributors: 55
Maintainers: 5 with merge privileges (CC'ed)
PyPI downloads:
Last week: 4.4K
Last month: 22K
All time: 150K
GitHub topics (ranked by number of stars):
Earth observation: #4
Satellite imagery: #4
Remote sensing: #7
Geospatial: #21

I believe we already meet most of the criteria for the highest level of membership in OSGeo:

Geospatial: yes
License: MIT <https://github.com/microsoft/torchgeo/blob/main/LICENSE>, including file headers
Contributions: 55 contributors and counting, contribution guide <https://torchgeo.readthedocs.io/en/stable/user/contributing.html>
Code of conduct: yes <https://github.com/microsoft/torchgeo/blob/main/CODE_OF_CONDUCT.md>
Community: active Slack workspace <https://join.slack.com/t/torchgeo/shared_invite/zt-22rse667m-eqtCeNW0yI000Tl4B~2PIw>
Users and contributors from both industry (Microsoft, Intel, Meta, IBM, Amazon) and academia (TUM, EPFL, ETH Zurich, Cambridge, Oxford, Stanford, Berkeley, CMU, MIT, U Washington)
GitHub: issue tracker <https://github.com/microsoft/torchgeo/issues>
Documentation: readthedocs <https://torchgeo.readthedocs.io/en/stable/>
Release procedure: wiki <https://github.com/microsoft/torchgeo/wiki/Releasing-Instructions>
CI: GitHub Actions <https://github.com/microsoft/torchgeo/tree/main/.github/workflows> with 100% test coverage <https://codecov.io/gh/microsoft/torchgeo>

We already have 10 stable releases and it's easy to install TorchGeo using pip, conda, or spack.

We think membership in the OSGeo Foundation would be beneficial to us in a number of ways. OSGeo represents the gold standard of open source software development in the geospatial community, and we would love your expertise to suggest ways we can grow our community of contributors. We would love to have tighter integration with other OSGeo projects like GDAL, working together to improve data loading performance. OSGeo provides a great opportunity to advertise our project on your website (and your foundation on our GitHub). We also have aspirations to present TorchGeo tutorials at as many remote sensing and machine learning conferences as possible, and could use funding to support travel for presenters.

Please let us know if TorchGeo fits your requirements for joining the OSGeo Foundation and how we can join. We would be interested in the highest level of membership possible, but are also happy to start off as a member of the OSGeo community and work our way up to OSGeo project or incubation.

Dr. Adam J. Stewart
Technical University of Munich
School of Engineering and Design
Data Science in Earth Observation
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/incubator/attachments/20231201/df7ece34/attachment.htm>