Jakub Nowosad
@jakubnowosad.com
Computational geographer. Associate Professor at AMU, Poznan, Poland. Co-author of http://r.geocompx.org, http://py.geocompx.org, and http://tmap.geocompx.org books. #rstats #rspatial #geocompx
https://jakubnowosad.com/
https://jakubnowosad.com/
Reposted by Jakub Nowosad
infos about more great talks at the GI forum here: www.uni-muenster.de/Geoinformati...
GI-Forum
GI-Forum neu
www.uni-muenster.de
November 4, 2025 at 8:19 AM
infos about more great talks at the GI forum here: www.uni-muenster.de/Geoinformati...
However, a random sample in the prediction area is rarely available in the case of ecological/environmental studies. So then, you need to use some cross-validation method, and the question of which method to use should depend on your prediction domain. (3/3)
October 31, 2025 at 8:21 AM
However, a random sample in the prediction area is rarely available in the case of ecological/environmental studies. So then, you need to use some cross-validation method, and the question of which method to use should depend on your prediction domain. (3/3)
I think that the main point of Wadoux et al. is that the design-based approach (random sampling in the prediction area) is always the best, and both random and spatial is not as good (I do not like the title of that paper, which I find too "clickbaity", and which is, in my opinion, misleading).(2/3)
October 31, 2025 at 8:20 AM
I think that the main point of Wadoux et al. is that the design-based approach (random sampling in the prediction area) is always the best, and both random and spatial is not as good (I do not like the title of that paper, which I find too "clickbaity", and which is, in my opinion, misleading).(2/3)
This a topic actively being discussed in my research group, and we are now completing a publication discussing this issue. I cannot answer fully in a short post, but my I would recommend you to read doi.org/10.5194/gmd-.... (1/3)
kNNDM CV: k-fold nearest-neighbour distance matching cross-validation for map accuracy estimation
Abstract. Random and spatial cross-validation (CV) methods are commonly used to evaluate machine-learning-based spatial prediction models, and the performance values obtained are often interpreted as ...
doi.org
October 31, 2025 at 8:19 AM
This a topic actively being discussed in my research group, and we are now completing a publication discussing this issue. I cannot answer fully in a short post, but my I would recommend you to read doi.org/10.5194/gmd-.... (1/3)
No, I do not know of any wrappers of rasterio warp...
October 29, 2025 at 3:59 PM
No, I do not know of any wrappers of rasterio warp...
Reposted by Jakub Nowosad
I'll write a post soon specifically on #rstats, but this week's post includes a full example of how SedonaDB can do this just as well in R as it can in Python!
October 16, 2025 at 3:03 PM
I'll write a post soon specifically on #rstats, but this week's post includes a full example of how SedonaDB can do this just as well in R as it can in Python!