Community Insight: Reproducible Processing Toolbox 🧰

In user interviews, we found that many researchers want to work in a graphical GIS environment and a Jupyter Notebook, and they want to come away with a unified, reproducible, and shareable product.

QGIS features an excellent processing toolbox of geospatial algorithms. This enables researchers to access the computational power of GDAL/OGR and other providers without needing to code. QGIS also provides a model builder which enables users to combine processing operations into a single reusable algorithm. Users can even export those algorithms to Python scripts!

ArcGIS provides similar geoprocessing and ModelBuilder offerings, and while our interviewees have described this as a valuable teaching tool, they’ve also commented that a major downside or deal-breaker is the requirement for a pricey (starting at $700/year at time of writing) ArcGIS license. Some practitioners described creating tools with the ArcGIS ModelBuilder and then rewriting them later as open source software packages! That’s a lot of work.

We aim to create a new workflow for JuypterGIS, providing an approachable graphical environment with immediate feedback for geospatial processing operations, but with tight integration with Jupyter Notebooks.