In comeetie/geohclust: Hierarchal clustering of geographic object with contiguity constraints

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

geohclust

geohclust offers two functions ?geohclust_poly and ?geohclust_graph that enable the clustering of spatial data such as polygons with a hclust type approach but taking advantages of contiguity constraints. The contiguity naturally create a sparsely connected graph that can be leveraged to speed-up the calculations and deal with more than 30000 polygons in seconds.

Installation

You can install the development version of geohclust from GitHub with:

# install.packages("devtools")
devtools::install_github("comeetie/geohclust")

Example

This is a basic example, we first prepare some spatial polygons data, here the results at the municipality level in one french department for the :

library(geohclust)
library(dplyr)
library(sf)
data("modesshare")
deplist = c(37)
dep = modesshare |> 
  filter(DEP %in% deplist) 

Do the clustering and use the classical function from ?hclust (?plot.hclust and ?cutree):

hc=geohclust_poly(dep)
plot(hc)
cutree(hc,k=30) |> head(20)

You may also use the ?geocutree function which build directly a spatial data.frame with the clustering results:

plot(geocutree(hc,k=30))

comeetie/geohclust documentation built on May 23, 2022, 6:30 a.m.