GeoJupyter core community meeting 2025-01-14

Meeting notes

The first monthly gathering of the GeoJupyter core community. Open to all!

Author

The GeoJupyter community

Published

January 14, 2025

GeoJupyter core community meeting 2025-01-14

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

    • Drive Let me know if you need access!
    • Checking with QuantStack for timing preference

Discussion

  1. 👋 Introductions
    1. Ciera: Thank you for showing up!!!
    2. Ciera: Introduce yourselves - shoot for 30 seconds. Ciera first, then popcorn! Matt will go last.
      1. Who are you? Your affiliation? Why are you excited about GeoJupyter?
    3. Matt: finance tech, then NSIDC, now here! So excited to learn from all of you ♥️
      1. Thank you all for meeting together!
      2. Expecting Fernando at 8:15.
    4. Matt: About GeoJupyter
      1. ☂️ An “umbrella effort” to improve user experiences working with geospatial (& temporal) data in JupyterLab.
        1. 🤩 Check out JupyterGIS by QuantStack, the first project in the GeoJupyter GitHub org
      2. 🚀 Mission (PRs welcome!)
        1. Reimagine geospatial interactive computing experiences for education, research, and industry.
        2. 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.
      3. 🥅 Goals (3 - 6 months; non-exhaustive 😉):
        1. Interview users: >12 completed. See blog post for early findings.
        2. Have one demoable / teachable use case for EGU - April 27, 2025.
        3. Develop and “bakeoff” additional use cases!
        4. Build relationships with early users (classrooms, workshops, great documentation and tutorials) and contributors (community programming, welcoming environment, …)
  2. 💬 How would you like to communicate?
    1. ✅ Things we have so far:
      1. ⚡ Real-time chat: Zulip. Please join!
      2. ⚡ Shared calendar: You can add it to your personal calendar by clicking “Add to Google Calendar” at the bottom of the calendar.
      3. ⚡ GeoJupyter.org website - open source, open a PR! anyone can contribute content, e.g. blog posts, to communicate with the community.
      4. ⚡ Community meetings - open, like Jupyter meetings
        1. Every 2nd Thursday of every month at 8AM PT / 4PM CET – thanks for being here ♥️
        2. 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.
      5. ⚡ Code of conduct: not published yet; starting with Project Jupyter’s. Be kind to others ♥️
    2. 👂 What else?
      1. Mailing list? Atom/RSS feed?
      2. Note-taking tool?
      3. Additional meetings or events to increase collaboration?
        1. JupyterGIS catchup! Add to shared calendar
      4. Shared drive! E.g. sharing presentations, videos, etc.
  3. 🗺️ Early strategy proposal
    1. 👥 One early primary audience
      1. 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.
      2. University educators and students. Strong interest from university communities (UC Berkeley, CU Boulder, Stanford, Clark).
    2. 🎬 One early use case
      1. Begin showing GeoJupyter rather than telling.
      2. It should be usable in a teaching scenario (workshop, tutorial, classroom).
      3. It should address a concrete user need or improve a user experience.
      4. Candidate use case: Geospatial debugging
        1. 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.
        2. How can we substitute JupyterGIS for QGIS in this workflow and eliminate unnecessary and frustrating steps? One line of Python, not counting import.
        3. QuantStack has already done 99.99…% of the work 🎉
      5. Fernando use case
        1. 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.
        2. QuantStack: Code generation exists for interacting with the JupyterGIS document itself.
        3. QuantStack: What if the JupyterGIS project file itself was a notebook? Discussed very briefly in https://github.com/geojupyter/jupytergis/issues/41
      6. …? Need ideas! 💡
  4. 📅 Upcoming events - will we see you?
    1. April 27, 2025: European Geosciences Union. Want to have something to show, not just tell!
    2. May 13-15, 2025: Matt and Fernando considering teaching a workshop @ CSDMS 2025 annual meeting at CU Boulder SEEC building.
    3. Week of May 19th, 2025: Potential hackathon event in Boulder, CO! 🥳
      1. Hybrid is a strong priority!
      2. Coding contributions to the core of GeoJupyter
    4. May 28-30, 2025: Matt and Fernando considering attending/teaching workshop @ ESIIL Innovation Summit at CU Boulder SEEC building
    5. July 7-13, 2025 - SciPy in Seattle
    6. JupyterCon - not ready for public announcement
    7. 2025 - PyData Paris, organized by QuantStack
    8. Summer 2025 - OSGeo conference
      1. March 26-29 - FOSSGIS in Münster (in German)
      2. FOSS4G Europe, July 14-20 in Bosnia and Herzegovina
    9. Quantstack attending:
      1. GeoPython https://2025.geopython.net/ in Feb. 2025
      2. Living Planet Symposium https://lps25.esa.int/ in June 2025
  5. 💪 What needs can the core team address right now?
    1. Generating candidates for new use cases! Please post on Zulip!
    2. Talk about and explore candidate use cases
    3. Explore .ipynb as JupyterGIS project format
      1. A mechanism to hedge against file format decisions
      2. Current format is structured JSON. Simple. Can extend the notebook format to include the same information in notebook metadata.
      3. Visualization concerns - tooling for modern (in-memory) formats less stable.
      4. JupyterGIS should support any data formats like QGIS does.
      5. What about bundling everything needed to make a map in the project file? Different from the QGIS paradigm, provides significant value!
        1. “Sidecar formats” - active discussions in Jupyter community <TODO: Link!>
    4. Community programming advice! What community programming would you suggest to increase collaboration and engagement at this stage? Please post on Zulip!
    5. Try out JupyterGIS! Make contributions! 🚀
    6. What else do you think we need? 👂
    7. …? 👂
  6. ☑️ Review action items
    1. 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)!
  7. Misc
    1. 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!