如何从人口普查形状文件(邮政编码级别)中删除所有小岛?

how to remove all the small islands from the Census Shapefile (zip code level)?

我已经下载了邮政编码级别的人口普查 shapefile,cb_2017_us_zcta510_500k.shp (https://www.census.gov/geo/maps-data/data/cbf/cbf_zcta.html)

我还下载了允许我添加相应 STATE 变量的映射文件 (https://www.census.gov/geo/maps-data/data/zcta_rel_download.html)

我将两者合并,得到:

library(sf)
library(dplyr)

big_df 

Simple feature collection with 44434 features and 2 fields
geometry type:  MULTIPOLYGON
dimension:      XY
bbox:           xmin: -176.6847 ymin: -14.37374 xmax: 145.8304 ymax: 71.34122
epsg (SRID):    4269
proj4string:    +proj=longlat +datum=NAD83 +no_defs
First 10 features:
   ZCTA5CE10 STATE                       geometry
1      35442     1 MULTIPOLYGON (((-88.25262 3...
2      35442     1 MULTIPOLYGON (((-88.25262 3...
3      35442     1 MULTIPOLYGON (((-88.25262 3...

现在,我尝试过滤所有的小岛和阿拉斯加:

remove_list <-  c("02", "15", "72", "66", "78", "60", "69",
"64", "68", "70", "74", "81", "84", "86", "87", "89", "71", "76",
"95", "79")



big_df %>% filter(!STATE %in% map(remove_list, as.integer)) %>% 
  tm_shape(.) + tm_polygons('pt_count',palette = "Reds", 
                            style = "quantile", n = 10, 
                            title = "counts") 

但我还是得到了一些小岛。

我在这里错过了什么? 谢谢!

这是获取美国大陆几何(国家概况)的方法:

library(raster)
library(sf)
library(dplyr)

us = getData('GADM', country='USA', level=0) %>%
  st_as_sf() %>%
  st_cast("POLYGON") %>%
  mutate(area = st_area(.)) %>%
  arrange(desc(area)) %>%
  slice(1) # mainland US should be the largest

然后您可以使用它来 运行 st_intersection(big_df, us) 仅提取 usbig_df 的部分。请注意,首先在 us 周围创建 st_bufferst_convex_hull 可能会有所回报,以确保您的 big_df 不会在其边界的某处被剪裁。