#rspatial

modtools.wordpress.com

Centering a map projection on the mapped region

The PROJ format for defining Coordinate Reference Systems (CRS) is becoming deprecated, as PROJ strings cannot accurately store modern datum transformations, coordinate epoch information, or complex axis order rules. However, that format is much easier and more practical for humans to type and read, and it is still often used as a legacy convenience. I, for one, use it often for centering a map projection on the mapped region, thus avoiding the distortion that occurs away from the projection center. For example, imagine I want to map New Zealand in a Lambert Azimuthal Equal-Area projection: library(terra) library(geodata) nz <- gadm(country = "New Zealand", level = 0, path = tempdir()) nz_prj <- project(nz, "+proj=laea") plot(nz_prj) The map is heavily distorted, because it is far away from the center of the projection, which is 0,0 by default. With the PROJ format, we can easily set the center wherever we want, e.g.: nz_centr <- crds(centroids(aggregate(nz))) nz_centr # 172.6104 -41.82592 nz_prj <- project(nz, "+proj=laea +lat_0=-41.8 +lon_0=172.6") # (this can be further automated to avoid the hard-coding, # but I keep it like this here for simplicity) plot(nz_prj) Much better! We now have an equal-area map of New Zealand without visibly distorting the shape, and without having to look for specific equal-area CRSs normally used over there (and again every time we need to map another region of the world). This works as long as we use the WGS84 ellipsoid rather than a local datum, as those are currently not supported when using PROJ. So, I’m still using this regularly, and apprehensive about PROJ being deprecated… If anyone knows of a non-PROJ way of easily defining user-centered projections like this, please let me know! ### Share this: * Share on X (Opens in new window) X * Share on Facebook (Opens in new window) Facebook * Like Loading... ### _Related_

jakubnowosad.com

Rethinking Validation for Spatial Machine Learning: Takeaways from the Talk – Thinking in spatial patterns

A summary of key points from my keynote and workshop at the Machine Learning for Earth Observation conference in Exeter (2026-06-22 and 2026-06-23).

jakubnowosad.com

Rethinking Validation for Spatial Machine Learning: Takeaways from the Talk – Thinking in spatial patterns

A summary of key points from my keynote and workshop at the Machine Learning for Earth Observation conference in Exeter (2026-06-22 and 2026-06-23).

github.com

GitHub - r-spatial/cesium: Cesium wrapper to R

Cesium wrapper to R. Contribute to r-spatial/cesium development by creating an account on GitHub.

cran.r-project.org

cartographr: Crafting Print-Ready Maps and Layered Visualizations

Simplifying the creation of print-ready maps, this package offers a user-friendly interface derived from 'ggplot2' for handling OpenStreetMap data. It streamlines the map-making process, allowing…

cran.r-project.org

cartographr: Crafting Print-Ready Maps and Layered Visualizations

Simplifying the creation of print-ready maps, this package offers a user-friendly interface derived from 'ggplot2' for handling OpenStreetMap data. It streamlines the map-making process, allowing users to focus on the story their maps tell. Transforming raw geospatial data into informative visualizations is made easy with simple features 'sf' geometries. Whether for urban planning, environmental studies, or impactful public presentations, this tool facilitates straightforward and effective map creation. Enhance the dissemination of spatial information with high-quality, narrative-driven visualizations!

fosstodon.org

Original post on fosstodon.org