Geographically Weighted Panel Regression

A modern, first implementation of Geographically Weighted Panel Regression (GWPR) for spatial panel data. The package provides a unified public API supporting Gaussian and binomial family models, within/pooling/random panel effects, three bandwidth search strategies (grid, Stochastic Gradient Descent, random), five kernel functions, and optional parallel execution via the 'future' framework. Diagnostic tools include spatial Moran's I, local F-test, Hausman test, and Lagrange Multiplier test.

GWPR.light

A modern, sf-first implementation of Geographically Weighted Panel Regression (GWPR) for spatial panel data. Version 1.0.0 provides a clean public API, three bandwidth search strategies (grid, SGD, random), Gaussian and binomial model families, and optional parallel execution via the future framework.

Authors

Chao Li [email protected]
Shunsuke Managi [email protected]

Installation

Install the released version from CRAN:

install.packages("GWPR.light")

Or install the development version from GitHub:

# install.packages("remotes")
remotes::install_github("MichaelChaoLi-cpu/GWPR.light")

Quick start (1.0.0 API)

The four public functions form the complete interface:

library(GWPR.light)
library(sf)

# --- Simulate a small spatial panel ---
set.seed(1)
pts <- sf::st_as_sf(
  data.frame(id = 1:6, X = c(0,1,2,0,1,2), Y = c(0,0,0,1,1,1)),
  coords = c("X", "Y"), crs = NA_integer_
)
dat <- data.frame(
  id   = rep(1:6, each = 4),
  time = rep(1:4, 6),
  x1   = rnorm(24),
  x2   = rnorm(24)
)
dat$y <- 1.5 * dat$x1 - 0.8 * dat$x2 + rnorm(24, sd = 0.3)

# --- Fit with a known bandwidth ---
fit <- fit_gwpr(
  y ~ x1 + x2, data = dat, spatial = pts,
  id = "id", time = "time",
  bandwidth = 2, model = "pooling", workers = 1
)
print(fit)

Automatic bandwidth search

bw <- select_bandwidth(
  y ~ x1 + x2, data = dat, spatial = pts,
  id = "id", time = "time",
  method  = "grid",
  control = list(lower = 0.5, upper = 3, step = 0.5),
  workers = 1
)
cat("Optimal bandwidth:", bw$best_bandwidth, "\n")

Diagnostics

diag_result <- diagnose_gwpr(fit, diagnostics = c("f_test", "hausman"))
print(diag_result)

Vignettes

• vignette("gwpr-introduction") — Getting started with the 1.0.0 API
• vignette("introduction_of_GWPR") — Legacy 0.2.x API reference

References

• Fotheringham, A., Brunsdon, C., Charlton, M. (2002). Geographically Weighted Regression. John Wiley & Sons. ISBN: 978-0-470-85525-6.
• Beenstock, M., Felsenstein, D. (2019). The Econometric Analysis of Non-Stationary Spatial Panel Data. Springer. ISBN: 978-3-030-03614-0.

Bug reports

https://github.com/MichaelChaoLi-cpu/GWPR.light/issues