Blas M. Benito
@blasbenito.com
🌍 Spatial Data Scientist | Team Lead | 🌱 AgTech | PhD in Computational Ecology + MSc in GIS
#rstats developer | geospatial engineering | soil microbiome and crop health mapping | product development
https://github.com/BlasBenito - www.blasbenito.com
#rstats developer | geospatial engineering | soil microbiome and crop health mapping | product development
https://github.com/BlasBenito - www.blasbenito.com
will do!
a brown and white guinea pig eating a sliced cucumber
ALT: a brown and white guinea pig eating a sliced cucumber
media.tenor.com
January 15, 2026 at 11:16 PM
will do!
It could have been a 12GB JSON!
January 15, 2026 at 5:58 PM
It could have been a 12GB JSON!
I didn’t know about {BH}, thanks for that!
My intuition is that spatial indexing only really pays off once the dataset is large enough that pairwise comparisons are no longer feasible. Below that threshold, the index construction and traversal overhead makes it slower than brute-force approaches.
My intuition is that spatial indexing only really pays off once the dataset is large enough that pairwise comparisons are no longer feasible. Below that threshold, the index construction and traversal overhead makes it slower than brute-force approaches.
Boost.Geometry (1.90.0)
The Boost.Geometry library provides geometric algorithms, primitives and spatial index.
www.boost.org
January 12, 2026 at 6:36 PM
I didn’t know about {BH}, thanks for that!
My intuition is that spatial indexing only really pays off once the dataset is large enough that pairwise comparisons are no longer feasible. Below that threshold, the index construction and traversal overhead makes it slower than brute-force approaches.
My intuition is that spatial indexing only really pays off once the dataset is large enough that pairwise comparisons are no longer feasible. Below that threshold, the index construction and traversal overhead makes it slower than brute-force approaches.
I've used it quite a lot and had a different experience. But I've been working on established code bases, so it never had to start from scratch.
January 12, 2026 at 4:37 PM
I've used it quite a lot and had a different experience. But I've been working on established code bases, so it never had to start from scratch.
Rcpp for the R <--> C++ infra, nothing more than that. The spatial index algo is written from scratch.
January 11, 2026 at 2:52 PM
Rcpp for the R <--> C++ infra, nothing more than that. The spatial index algo is written from scratch.
In the context of spatialRF, thinning is used to define the centers of contiguous training folds used in spatial cross-validation (shown as blue dots in the figure).
These ensure that the training data represents the spatial correlation structure of the full dataset.
These ensure that the training data represents the spatial correlation structure of the full dataset.
January 11, 2026 at 9:42 AM
In the context of spatialRF, thinning is used to define the centers of contiguous training folds used in spatial cross-validation (shown as blue dots in the figure).
These ensure that the training data represents the spatial correlation structure of the full dataset.
These ensure that the training data represents the spatial correlation structure of the full dataset.
In case "thinning" doesn't ring a bell:
spatialRF::thinning() controls spatial clustering in point data to mitigate spatial autocorrelation and sampling bias.
The ugly figure shows the before and after of an extreme thinning run with a distance of 5 degrees on a global dataset with 30k points.
spatialRF::thinning() controls spatial clustering in point data to mitigate spatial autocorrelation and sampling bias.
The ugly figure shows the before and after of an extreme thinning run with a distance of 5 degrees on a global dataset with 30k points.
January 11, 2026 at 9:17 AM
In case "thinning" doesn't ring a bell:
spatialRF::thinning() controls spatial clustering in point data to mitigate spatial autocorrelation and sampling bias.
The ugly figure shows the before and after of an extreme thinning run with a distance of 5 degrees on a global dataset with 30k points.
spatialRF::thinning() controls spatial clustering in point data to mitigate spatial autocorrelation and sampling bias.
The ugly figure shows the before and after of an extreme thinning run with a distance of 5 degrees on a global dataset with 30k points.
TIL Linus Torvalds doesn't know Python. I feel better about myself now!
January 11, 2026 at 12:19 AM
TIL Linus Torvalds doesn't know Python. I feel better about myself now!
I've been using it to manage my obsidian vault for a while. It's pretty good at it!
January 10, 2026 at 7:08 PM
I've been using it to manage my obsidian vault for a while. It's pretty good at it!
Messy commit messages aren't an issue, just part of the process IMHO.
However, if we are talking about portfolio projects, then having cleanly squashed commits might suggest some semblance of a well organized development process. That would be a lie, but a clean one!
However, if we are talking about portfolio projects, then having cleanly squashed commits might suggest some semblance of a well organized development process. That would be a lie, but a clean one!
January 9, 2026 at 7:39 AM
Messy commit messages aren't an issue, just part of the process IMHO.
However, if we are talking about portfolio projects, then having cleanly squashed commits might suggest some semblance of a well organized development process. That would be a lie, but a clean one!
However, if we are talking about portfolio projects, then having cleanly squashed commits might suggest some semblance of a well organized development process. That would be a lie, but a clean one!
No no, loops are cool!
I imagine your for loops are doing some recursive operation and are not embarrassingly parallel? Because my first though was "parallelize the loops and run them sequentially", but recursive operations don't fit with that.
I imagine your for loops are doing some recursive operation and are not embarrassingly parallel? Because my first though was "parallelize the loops and run them sequentially", but recursive operations don't fit with that.
January 8, 2026 at 12:05 PM
No no, loops are cool!
I imagine your for loops are doing some recursive operation and are not embarrassingly parallel? Because my first though was "parallelize the loops and run them sequentially", but recursive operations don't fit with that.
I imagine your for loops are doing some recursive operation and are not embarrassingly parallel? Because my first though was "parallelize the loops and run them sequentially", but recursive operations don't fit with that.