GeoJupyter core community meeting (2025-01-14)¶
GeoJupyter handy links:
Attendees¶
Ciera Martinez / DSE
Matt Fisher / DSE
Qiusheng Wu / UTK
Maryam Vareth / BIDS
Martin Renou / QuantStack
Tammy Woodard / Clark CGA
Greg Mooney / QuantStack
Nicolas Brichet / QuantStack
Kyle Barron / Development Seed
Fernando Pérez / DSE/BIDS/Berkeley
Action items¶
All: Please read the GeoJupyter announcement blog post!
All: Please introduce yourself in our Zulip “welcome” thread!
All: Consider upcoming events below – will we see you there?
All: Consider “What needs can the core team address right now?” below. Can you help with any of these things?
All: How are we doing? What do you think of our goals and needs? Have we missed anything? For any feedback, please create a new topic on Zulip, or reach out to Matt Fisher in whatever medium you prefer!
~~Matt: Create shared drive~~
Drive Let me know if you need access!
~~Matt: Schedule JupyterGIS catch-up meeting on shared calendar~~
Checking with QuantStack for timing preference
~~Matt: Create a bucket for use case ideas~~
Discussion¶
👋 Introductions
Ciera: Thank you for showing up!!!
Ciera: Introduce yourselves - shoot for 30 seconds. Ciera first, then popcorn! Matt will go last.
Who are you? Your affiliation? Why are you excited about GeoJupyter?
Matt: finance tech, then NSIDC, now here! So excited to learn from all of you ♥️
Thank you all for meeting together!
Expecting Fernando at 8:15.
Matt: About GeoJupyter
☂️ An “umbrella effort” to improve user experiences working with geospatial (& temporal) data in JupyterLab.
🤩 Check out JupyterGIS by QuantStack, the first project in the GeoJupyter GitHub org
🚀 Mission (PRs welcome!)
Reimagine geospatial interactive computing experiences for education, research, and industry.
Combine the approachability and playfulness of desktop GIS tools, the flexibility, efficiency, and reproducibility of coding-driven GIS methods, and the collaborative and storytelling power of Jupyter to enable more researchers, educators, and learners to confidently engage with geospatial data.
🥅 Goals (3 - 6 months; non-exhaustive 😉):
Interview users: >12 completed. See blog post for early findings.
Have one demoable / teachable use case for EGU - April 27, 2025.
Develop and “bakeoff” additional use cases!
Build relationships with early users (classrooms, workshops, great documentation and tutorials) and contributors (community programming, welcoming environment, …)
💬 How would you like to communicate?
✅ Things we have so far:
⚡ Real-time chat: Zulip. Please join!
⚡ Shared calendar: You can add it to your personal calendar by clicking “Add to Google Calendar” at the bottom of the calendar.
⚡ GeoJupyter.org website - open source, open a PR! anyone can contribute content, e.g. blog posts, to communicate with the community.
⚡ Community meetings - open, like Jupyter meetings
Every 2nd Thursday of every month at 8AM PT / 4PM CET – thanks for being here ♥️
Post here on Zulip (or DM Matt) if this time is a struggle for you and we’ll attempt to find a better one or add a new time.
⚡ Code of conduct: not published yet; starting with Project Jupyter’s. Be kind to others ♥️
👂 What else?
Mailing list? Atom/RSS feed?
Note-taking tool?
Additional meetings or events to increase collaboration?
JupyterGIS catchup! Add to shared calendar
Shared drive! E.g. sharing presentations, videos, etc.
🗺️ Early strategy proposal
👥 One early primary audience
We can’t be pulled in too many directions at this stage; we want to build some thing(s) that meet(s) a specific need so we can get feedback.
University educators and students. Strong interest from university communities (UC Berkeley, CU Boulder, Stanford, Clark).
🎬 One early use case
Begin showing GeoJupyter rather than telling.
It should be usable in a teaching scenario (workshop, tutorial, classroom).
It should address a concrete user need or improve a user experience.
Candidate use case: Geospatial debugging
A very common workflow: Build an analysis or tutorial in a Jupyter Notebook, but constantly writing out intermediate files (transferring between computers if necessary) and “bouncing” between the notebook and QGIS to validate the data. High friction, high cognitive load, disruptive to screen sharing when teaching or collaborating.
How can we substitute JupyterGIS for QGIS in this workflow and eliminate unnecessary and frustrating steps? One line of Python, not counting import.
QuantStack has already done 99.99…% of the work 🎉
Fernando use case
Open a notebook from JupyterGIS session, pre-populate some cells: Markdown header cell with some details about data; code cells containing an object for each data layer in JGIS and necessary python imports to operate on.
QuantStack: Code generation exists for interacting with the JupyterGIS document itself.
QuantStack: What if the JupyterGIS project file itself was a notebook? Discussed very briefly in geojupyter
/jupytergis #41
…? Need ideas! 💡
📅 Upcoming events - will we see you?
April 27, 2025: European Geosciences Union. Want to have something to show, not just tell!
May 13-15, 2025: Matt and Fernando considering teaching a workshop @ CSDMS 2025 annual meeting at CU Boulder SEEC building.
Week of May 19th, 2025: Potential hackathon event in Boulder, CO! 🥳
Hybrid is a strong priority!
Coding contributions to the core of GeoJupyter
May 28-30, 2025: Matt and Fernando considering attending/teaching workshop @ ESIIL Innovation Summit at CU Boulder SEEC building
July 7-13, 2025 - SciPy in Seattle
JupyterCon - not ready for public announcement
2025 - PyData Paris, organized by QuantStack
Summer 2025 - OSGeo conference
March 26-29 - FOSSGIS in Münster (in German)
FOSS4G Europe, July 14-20 in Bosnia and Herzegovina
Quantstack attending:
GeoPython https://
2025 .geopython .net/ in Feb. 2025 Living Planet Symposium https://
lps25 .esa .int/ in June 2025
💪 What needs can the core team address right now?
Generating candidates for new use cases! Please post on Zulip!
Talk about and explore candidate use cases
Explore .ipynb as JupyterGIS project format
A mechanism to hedge against file format decisions
Current format is structured JSON. Simple. Can extend the notebook format to include the same information in notebook metadata.
Visualization concerns - tooling for modern (in-memory) formats less stable.
JupyterGIS should support any data formats like QGIS does.
What about bundling everything needed to make a map in the project file? Different from the QGIS paradigm, provides significant value!
“Sidecar formats” - active discussions in Jupyter community <TODO: Link!>
Community programming advice! What community programming would you suggest to increase collaboration and engagement at this stage? Please post on Zulip!
Try out JupyterGIS! Make contributions! 🚀
What else do you think we need? 👂
…? 👂
☑️ Review action items
If you have more you’d like to talk about, post a new topic on Zulip, reach out to Matt Fisher on Zulip, or book a meeting (doesn’t have to be an interview)!
Misc
Fernando @ AGU: Attended 3 hour GHG workshop. Part in notebooks, part in QGIS. Immediately illustrated why GeoJupyter is important. QGIS in JupyterHub with VNC on a crowded shared wifi is a miserable experience! People began falling behind the instructor. In a GUI workflow, if you miss a step in a workshop, you can’t go back and see what button you were supposed to press!