Jakub Nowosad is a computational geographer and Associate Professor at Adam Mickiewicz University, also serving as a Visiting Scientist at the University of Münster. He develops open-source tools and spatial methods for reproducible, scalable environmental and ecological analysis, and co-authors Geocomputation with R and Geocomputation with Python.
42 articles from this blog
Explores the application of Graph Neural Networks, embeddings, and foundation models to spatial data science, with practical examples in R.
Summary of a talk on using R for geospatial predictive mapping, covering methods like Kriging and Random Forests, and tools for evaluating prediction reliability.
A researcher shares progress on the PRISM project, an MSCA-PF grant focused on validating spatial patterns in machine learning for remote sensing.
A guide to methods and R implementations for comparing spatial patterns in raster data, covering continuous and categorical data for overlapping and arbitrary regions.
A technical guide comparing spatial patterns in categorical raster data using R, focusing on landscape metrics and code examples.
A technical guide comparing spatial patterns in categorical raster data using R, focusing on land cover change analysis.
A technical guide on comparing spatial patterns in continuous raster data using R, focusing on methods for arbitrary regions.
A technical guide comparing spatial patterns in continuous raster data for overlapping regions using R, focusing on NDVI data analysis.
A technical guide to methods for comparing spatial patterns in raster data, focusing on analysis using R for both continuous and categorical data.
Using simulated annealing to optimize parameters (B and J) in a spatial kinetic Ising model for simulating spatial patterns.
Explains how to use the spatial kinetic Ising model in R to simulate changes in binary spatial patterns, like land cover.
Explains how to extract and analyze spatial patterns from categorical raster data using the 'motif' R package and information theory metrics.
Explores unsupervised methods (k-means and SKATER) for merging homogeneous supercells into larger regions for geospatial data analysis.
Explains the Supercells algorithm for generating superpixels to improve segmentation of geospatial and satellite imagery.
Final blog post in a series on the 'motif' R package, summarizing pattern-based spatial analysis methods and future research directions.
A technical guide on clustering similar spatial patterns from raster data using R, focusing on land cover and landform analysis in Africa.
A technical guide on using R and spatial analysis to quantify changes in land cover patterns in the Amazon between 1992 and 2018.
A technical tutorial on using R and geospatial analysis to find areas with similar topography to a query region, focusing on spatial pattern matching.
Explains spatial signatures and co-occurrence matrices for analyzing patterns in categorical raster data like land cover.
An introduction to pattern-based spatial analysis in R using the 'motif' package for analyzing categorical raster data like land cover maps.
