Planned features

• Support spatial regression models from spreg and the PySAL stack (Rey and Anselin, 2009).

• Add a method to plot scalograms, i.e., plotting how the computed spatial predictors respond to changes in scale, which can reveal scale thresholds that maximize landscape heterogeneity (Pasher et al., 2013) (and therefore the variance of the spatial predictors that act as independent variables).

• Implement algorithms to sample locations for field data collection based on landscape heterogeneity (Bowler et al., 2022).

• Add methods to assess the “area of applicability” (Meyer and Pebesma, 2021) of the models (based on the latent space defined by the spatial predictors) as well as the “risk of spatial extrapolation” (Gutzwiller and Serno, 2023).