PlanX GeoStats Lab
PlanX spatial analysis lab for QGIS planning workflows: data preparation, urban pattern scanning, hot spots, centers, exploratory models and scenarios, machine learning and explainable AI, and geostatistical interpolation.
PlanX GeoStats Lab brings advanced spatial statistics and machine learning tools to the QGIS Processing Toolbox with a planning-analysis workflow designed for PlanX: Setup and Diagnostics, Data Preparation and Neighborhoods, Urban Pattern Scan, Hot Spots and Spatial Outliers, Centers Direction and Dispersion, Models and Scenarios, Machine Learning and Explainable AI, and Interpolation and Geostatistics. The geostatistics group adds the classical two-step workflow: Empirical Semivariogram and Model Fitting estimates how spatial similarity decays with distance and fits a theoretical model to it, and Ordinary / Universal Kriging Interpolation builds a surface from those fitted parameters while writing kriging's own analytically derived variance as a second raster, so every prediction arrives with the model's stated uncertainty. The group also adds Empirical Bayesian Kriging, which fits the variogram itself and reports an uncertainty that accounts for having fitted it, Indicator Kriging, which answers how likely each location is to be above a level you name without assuming any distribution for the field, and an Interpolation Model Comparison that scores the interpolators against one another under a single shared validation protocol. The model workflow includes OLS, Generalized Linear Regression, Spatial Lag Regression, Spatial Error Regression, Spatial Durbin, Spatial Regime Regression, Exploratory Regression, Quantile Regression, GWR, MGWR, Eigenvector Spatial Filtering, model comparison and sensitivity testing. The Machine Learning group adds Random Forest, Extra Trees, Gradient Boosting (scikit-learn/XGBoost/LightGBM/CatBoost), SVM, Neural Network, and TabPFN (2025 tabular foundation model) regression/classification; spatial k-fold cross-validation (K-Means block or kNNDM); and explainability tools including SHAP (global/spatial map/local), the exact-contribution Explainable Boosting Machine, DiCE counterfactual explanations, and distribution-free Conformal Prediction Intervals. Optional GeoStats Python libraries are managed only through Processing Toolbox tools under 00 | Setup and Diagnostics: GeoStats Library Status for non-installing diagnostics and Install / Update GeoStats Libraries for explicit, user-approved pip installation. Developed with feedback from educational workflows at Dokuz Eylul University, Department of City and Regional Planning. Online User Manual & Documentation: https://geophilo.com/planx_geostats/GEOSTATS_REFERENCE_MANUAL.html | If you find this plugin helpful, please consider starring the repository on GitLab (https://gitlab.com/geophilo1/planx_geostats)!