Spatial regression models- LI post

How do we model the geography of real-world problems — from disease patterns to environmental change?

Researchers from across Africa recently gathered in Bohicon, Benin, for an intensive four-day training on spatial regression models and applied spatial statistics, hosted by the Laboratory of Biomathematics and Forest Estimation (LABEF) at the University of Abomey-Calavi.

Supported by the German Foreign Office through the SEMCA Research Hub, the training was led by Professor Romain Glèlè Kakaïand Dr Kolawolé Valère SALAKO, with support from postdoctoral researchers and PhD students. Participants explored how spatial thinking and spatial econometric models can be used to better understand complex phenomena in public health, agronomy, and environmental science. The programme covered topics such as spatial autocorrelation, spatial weight matrices, and advanced spatial regression approaches including Spatial Lag Models, Spatial Error Models, Spatial Durbin Models, and Geographically Weighted Regression. Through hands-on sessions in R, participants applied spatial modelling techniques using tools such as spdep, sf, spgwr, mgwrsar, ggplot2, and tmap to analyse and visualise spatial data.

The training brought together 40 in-person participants and more than 200 online participants from 16 African countries, creating a vibrant space for learning, exchange, and collaboration. Notable participants included Souand Tahi, Midokpè Merveille Essetcheou, Surya N. L AHAMIDE, KANGELA MATAZI Alain, kandala desire, Elysee Tshiama Kabongo, and Marcel DONOU.

Why this matters
Spatial statistical methods are increasingly essential for understanding geographically distributed challenges, from disease transmission to environmental change. Strengthening these skills across African research communities helps ensure that scientists on the continent are leading the development and application of advanced analytical methods.