Over the course of two lively days, 25 developers, researchers, and enthusiasts from a range of backgrounds came together in Berkeley for the GeoJupyter Hackathon. The event was marked by the energy and momentum generated by people working side by side. Participants convened at the AI Futures Lab on day one, which provided a space made for collaboration, with views and natural light that helped keep the environment comfortable and focused, supporting open discussion and steady progress. The Eric and Wendy Schmidt Center for Data Science and Environment (DSE) and the Berkeley Institute for Data Science (BIDS) co-hosted the event, providing equipment, meals, and problem-solving support to keep everyone energized.
Participants arrived with different areas of expertise, from geospatial analysis to software development, and varying levels of familiarity with JupyterGIS, a flagship project within GeoJupyter that has features including a collaborative GIS environment, QGIS file support, and interactive maps. This mix of skills and experiences proved valuable for collaborative problem-solving. The first day began with conversations rather than coding, sharing visions for what the group hoped to build, defining the kinds of user interactions that would make the tools more intuitive, and identifying potential roadblocks. By airing potential difficulties early, the teams could brainstorm ways to minimize them.
Photo: A group of on-site and remote attendees getting started on their project.
One of the goals of the Hackathon was to make the tools reproducible and shareable, but discussions extended beyond the technical level. Participants considered the kinds of problems the project could address and where it could add real value, considering both immediate needs and long-term possibilities. These conversations encouraged creative thinking while keeping expectations realistic. Rather than locking into one narrow definition, the project could be seen as a growing ecosystem shaped by the needs of its community.
As the teams began working, the in-person setting provided opportunities to move around the rooms, lean over each other’s laptops, sketch ideas on whiteboards, and check in on the progress of other groups. Laughter and focused problem-solving conversations filled the space. Collaboration remained fluid, with participants shifting between tasks and forming subgroups based on immediate challenges rather than fixed roles.
The collaborative and focused environment was noted throughout the event. Matt Fisher, Research Software Engineer and Community Manager at DSE, who facilitated the hackathon, explained, “This was my first time hosting an in-person hackathon, and while it took a lot of effort, I came away energized to do it again. It was intensely rewarding bringing together intellects and personalities that I admire to work together on problems that matter to real people. You challenged my assumptions, and I learned so much from you, and I appreciate that deeply!” That sense carried through both days, reflecting that open-source projects thrive not only on code, but also on community.
Photo: Matt Fisher presenting slides to the group about GeoJupyter and JupyterGIS.
At the end of the two days, participants had built relationships, exchanged new ideas, and planned how to continue the work. The hybrid format allowed remote voices to join, but the consensus in the room was clear: a central space for face-to-face exchange accelerated collaboration, strengthened relationships, and gave the project new momentum. The GeoJupyter hackathon demonstrated that progress in technology is tied to human collaboration.
The work will continue in the coming weeks, and the lessons and collaborative spirit of those two days will guide future efforts. DSE’s organization and facilitation made this special event possible. A full summary of the outcomes is available here: In-person Hackathon and Design Dialog 2025.