Predictive land value modelling in Guatemala City using a geo-statistical approach and Space Syntax

Specifically, the datasets were produced and prepared with the purpose of training a geostatistical model and construct a residential land value map for Guatemala City using a predictive approach. The file "gc_log" contains a set of spatial points representing parcel centroids of the observations that were used to train a regression-kriging model.

A spatialized variable selection was implemented in order to retain only important variables. The "grid_prediction.csv" contains a set of spatial points representing the centroids of an hexagonal tessellation containing values for all the required predictors of land value. The "research_code.R" file contains the scripts developed to implement the spatial variable selection, train the regression-kriging model and construct the land value map.