Tools for Statistical Inference with Geo-Coded Data R package SpatialInference version 0.1.0 Fast computation of Conley (1999) <doi:10.1016/S0304-4076(98)00084-0> spatial heteroskedasticity and autocorrelation consistent (HAC) standard errors for linear regression models with geo-coded data, with a fast C++ implementation by Christensen, Hartman, and Samii (2021) <doi:10.1017/S0020818321000187>. Performance-critical distance calculations, kernel weighting, and variance component accumulation are implemented in C++ via 'Rcpp' and 'RcppArmadillo'. Includes tools for estimating the spatial correlation range from covariograms and correlograms following the bandwidth selection method proposed in Lehner (2026) <doi:10.48550/arXiv.2603.03997>, and diagnostic visualizations for bandwidth selection. Software Depends: R (>= 4.1.0) Imports: Rcpp, sf, data.table, magrittr, stats, tibble LinkingTo: Rcpp, RcppArmadillo Suggests: lfe, fixest, dplyr, stringr, spdep, ncf, gstat, sandwich, ggplot2, modelsummary, knitr, rmarkdown, geosphere, testthat (>= 3.0.0) Alexander Lehner <alehner@worldbank.org> Comprehensive R Archive Network (CRAN) Alexander Lehner [aut, cre] (ORCID: <https://orcid.org/0000-0001-5885-5966>) GPL (>= 3) 2026-03-25 application/tgz https://CRAN.R-project.org/package=SpatialInference doi:10.32614/CRAN.package.SpatialInference