A world still to be mapped: reflections on geocomputation in R: takeaways from the talk and workshop at UseR! 2026

At useR! 2026 in Warsaw, I contributed to a workshop on Geocomputation with R (July 6, 2026) and gave a keynote entitled A world still to be mapped: reflections on geocomputation in R (July 8, 2026). Both were opportunities to consider what the R spatial ecosystem already makes possible, its strengths, limitations, and what remains to be done.

Keynote slides: https://jakubnowosad.com/user2026/

Workshop materials: https://github.com/geocompx/user26

Workshop

The workshop, led by Jannes Muenchow and based on the second edition of Geocomputation with R, introduced some of the core building blocks of spatial work in R. It covered vector data with the sf package, raster data with the terra package, and spatial data visualization with the tmap package. I presented the visualization part, showing the main ingredients of the mapping workflow and demonstrating how maps can be used to explore geographic data and communicate analytical results.

These tools represent a mature and increasingly interoperable ecosystem. They make it possible to import, manipulate, analyze, and visualize geographic data while drawing on R’s wider capabilities for data processing and statistics. The workshop combined presentations, live coding, and exercises to provide a practical starting point for working with these tools. You can try the exercises yourself by following the instructions in the workshop materials.

Talk

At UseR! 2026, I spoke about geocomputation in R through the lens of my work as a computational geographer. I started by showing that geography is not only about drawing maps or attaching coordinates to data. For me, it is a way of asking questions about what is happening, where it is happening, why it is happening there, and how those patterns change over time. To answer those questions, I use a combination of spatial data, statistical models, and visualization.

I also reflected on how far the R spatial ecosystem has come. Over the last few decades, R has grown into a strong environment for spatial analysis, with tools for vector and raster data, visualization, reproducibility, and integration with broader geospatial infrastructure. At the same time, some of the most important gaps are no longer just technical. We can keep building better packages, but that alone will not solve problems like weak validation practices or misleading maps – the challenge is not only to make better maps, but to make predictions and maps that we can trust. For these purposes, I discussed ideas such as prediction-domain adaptive evaluation in the context of spatial machine learning and the importance of proper color palettes and projection choices in the context of spatial visualization.

My conclusion was that the future of geocomputation in R depends as much on community as on code. The R spatial ecosystem is strong in scientific rigor, geospatial infrastructure, visualization, and reproducible workflows, but these strengths depend on an active community. Good spatial work still requires data, domain knowledge, careful evaluation, and honest communication. We should, therefore, continue building tools while also supporting people who maintain software, share examples and expertise, mentor others, and connect methods to real-world problems.

@online{nowosad2026,
  author = {Nowosad, Jakub},
  title = {A World Still to Be Mapped: Reflections on Geocomputation in
    {R:} Takeaways from the Talk and Workshop at {UseR!} 2026},
  date = {2026-07-21},
  url = {https://jakubnowosad.com/posts/2026-07-21-user/},
  langid = {en}
}