英文:
How to calculate near neighbours in sf object with polygon geometry in R?
问题
我有德国邮政编码的多边形形状数据。对于这些邮政编码多边形，我想计算各种最近邻居度量。我已经看到使用sp包的方法（使用coordinates()，例如knearneigh(coordinates(GER), k = 4)）。我选择在R中使用sf空间对象，但对如何在这里实现邻居感到困惑。谢谢
library(sf)
library(dplyr)
library(leaflet)
URL <- "https://downloads.suche-postleitzahl.org/v2/public/plz-5stellig.shp.zip"
# 使用GDAL虚拟文件系统从远程URL加载压缩的shapefile
GER_postcode <- paste0("/vsizip//vsicurl/", URL) %>% read_sf()
# 使用giscoR获取德国国界轮廓
GER_outline <- giscoR::gisco_get_countries(country = "DE")
# 子样本
GER_postcode_subsample <- GER_postcode %>% filter(substr(plz, 1, 1) %in% c(0, 1, 7))
# sf数据框的k个最近邻居
英文:
I have polygon shape data for German postcodes. For those postcode polygon I like to calculate various nearest neighbour measures. I have seen that procedures working with the sp package (using coordinates(), like knearneigh(coordinates(GER), k = 4)). I opt for sf spatial objects in R and are confused on how to implement neighbours here. Thank you
library(sf)
library(dplyr)
library(leaflet)
URL <- "https://downloads.suche-postleitzahl.org/v2/public/plz-5stellig.shp.zip"
# use GDAL virtual file systems to load zipped shapefile from remote url
GER_postcode <- paste0("/vsizip//vsicurl/", URL) %>% read_sf()
# country outline from giscoR
GER_outline <- giscoR::gisco_get_countries(country = "DE")
# subsample
GER_postcode_subsample <- GER_postcode %>% filter(substr(plz, 1, 1) %in% c(0, 1, 7))
# k nearest neighbours for sf dataframe
答案1
得分: 0
我在spdep包中找到了答案，其中包含了名为poly2nb()的函数。不知道为什么我之前没有找到它。
library(spdep)
queens <- poly2nb(GER_postcode_subsample,
queen = TRUE, # 单个共享边界点符合邻近条件
snap = 1) # 我们将1米内的点视为“相邻”
summary(queens)
邻居列表对象：
区域数量：2624
非零链接数量：13698
非零权重百分比：0.1989434
平均链接数量：5.220274
没有链接的9个区域：
275 284 554 616 922 947 1889 2328 2329
链接数量分布：
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
9 28 117 307 513 574 468 310 168 78 28 11 2 10 1
28个最少连接的区域：
112 297 298 474 529 809 843 852 896 917 921 946 951 1027 1050 1147 1524 1687 1884 2068 2271 2291 2314 2327 2343 2367 2368 2509，每个区域有1个链接
1个最连接的区域：
1349，有14个链接
英文:
I found the answer in the spdep package which contains the speaking function poly2nb(). Don't know why I havn't found this earlier.
library(spdep)
queens <- poly2nb(GER_postcode_subsample,
queen = TRUE, # a single shared boundary point meets the contiguity condition
snap = 1) # we consider points in 1m distance as 'touching'
summary(queens)
Neighbour list object:
Number of regions: 2624
Number of nonzero links: 13698
Percentage nonzero weights: 0.1989434
Average number of links: 5.220274
9 regions with no links:
275 284 554 616 922 947 1889 2328 2329
Link number distribution:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
9 28 117 307 513 574 468 310 168 78 28 11 2 10 1
28 least connected regions:
112 297 298 474 529 809 843 852 896 917 921 946 951 1027 1050 1147 1524 1687 1884 2068 2271 2291 2314 2327 2343 2367 2368 2509 with 1 link
1 most connected region:
1349 with 14 links
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论