2026-11-03 –, Bondi
The Jupyter architecture is more powerful than you think. We introduce and demonstrate GeoJupyter, a community effort for reimagining geospatial workflows and tools beyond the map through Notebooks, real-time collaboration, and AI integration.
In this talk, we discuss how GeoJupyter leverages the architecture of Project Jupyter to build tools that bring interactive computational exploration to geospatial workflows, while integrating with the established and rich open source geospatial ecosystem.
Geospatial data, from Earth system models to satellite and field observations, continues to grow in size, complexity, and day-to-day importance, making open and fluid interoperability increasingly crucial. This interoperability is both more challenging and impactful in workflows that combine local and cloud resources.
The focus of GeoJupyter is to explore how the architecture and infrastructure of Project Jupyter can best contribute to this challenge, as part of the existing ecosystem of open geospatial tools for research, education and industry. Rather than replacing traditional GIS, it strives for a cooperative and complementary approach, emphasizing user experience, computational analysis, reproducibility, and collaboration. From a community perspective, GeoJupyter is an open initiative that welcomes contributions from all, with a goal of democratizing access to the computational analysis of data about our world.
The vision for GeoJupyter is different from map-centric GIS tools. For GeoJupyter, maps are critically important, but they aren't the only lens, or even the main one; we are instead building components that support the exploration and analysis of geospatial data using the full suite of tools and APIs in Jupyter while complementing existing projects like OSGeo, QGIS, Pangeo, the Open Geospatial Consortium (OGC), and Cloud Native Geospatial (CNG). Map-based workflows are supported through the JupyterGIS library (development led by QuantStack), which can be used as needed alongside Notebooks, scripts, and other tools.
In our presentation, we first step back from geospatial data to explain the core ideas of Jupyter itself: Notebooks, kernels, network protocols, collaboration, publishing, and AI integration. We then discuss the GeoJupyter vision to use these building blocks to provide a flexible and exploratory playground where users of geospatial models and data can combine multiple kinds of data (e.g. raster images, vector layers, tabular sources, n-dimensional dense arrays from numerical models, etc.), run computational models, collaborate in real time, conduct statistical analyses, visualize data, and get assistance as relevant by AI tools that can rely on commercial services or run on private, user-controlled models. After our overview, we present live demos of the tools we describe, integrating map-driven geospatial exploration with computational tools, AI support, and reproducible publishing.
Fernando Pérez is an Associate Professor in Statistics at UC Berkeley, faculty Scientist at Lawrence Berkeley National Laboratory, faculty director of the Berkeley Institute for Data Science (BIDS), and co-director of the Eric and Wendy Schmidt Data Science & Environment Center (DSE). His research focuses on creating tools for computational research across disciplines, with an emphasis on interactive computing, open science and reproducible research. He is particularly interested in the intersection of physics, data and AI in earth science, to address critical problems such as climate change and environmental protection, as well as supporting open source and AI in healthcare. He created IPython while a graduate student in 2001 and co-founded its successor, Project Jupyter. Also an entrepreneur, he is a co-founder and CTO of AIMATX, a company that builds tools for AI-driven design of novel materials.
Pérez co-founded the non-profit 2i2c initiative and the NumFOCUS Foundation. He is the recipient of the 2024 Exceptional Public Service Medal from NASA, the 2024 "Champions of Open Science" recognition from the White House Office of Science and Technology Policy, the 2017 ACM Software Systems Award, and the 2012 Award for the Advancement of Free Software from the Free Software Foundation. He is a Kavli Frontiers Fellow of the US National Academy of Sciences, and holds a PhD in physics from the University of Colorado, Boulder.
