如何使用KML多边形过滤地理坐标的数据框？

Question

我有一个跨越整个地球的数据点CSV（the US Geological Service's earthquake feed），我想只过滤美国境内的地震＆＃39;边界。

我从美国人口普查局撤下的KML文件：

https://www.census.gov/geo/maps-data/data/kml/kml_nation.html

在R中，rgdal库可以加载KML文件：

library(rgdal)
kml = readOGR("kmls/cb_2014_us_nation_20m.kml", 'cb_2014_us_nation_20m')

如何使用dplyr / plyr /等。过滤data.frame（地理数据的列是latitude和longitude）仅针对属于KML指定的边界内的行？

编辑，回复后：

以下是@hrbrmstr's answer用于快速展示的内容：

library(sp)
library(rgdal)
# download earthquakes
url <- "http://earthquake.usgs.gov/earthquakes/feed/v1.0/summary/all_week.csv"
fil <- "all_week.csv"
if (!file.exists(fil)) download.file(url, fil)
quakes <- read.csv("all_week.csv", stringsAsFactors=FALSE)
# create spatial object
sp::coordinates(quakes) <- ~longitude+latitude

# download nation KML
url <- "http://www2.census.gov/geo/tiger/GENZ2014/kml/cb_2014_us_nation_20m.zip"
fil <- "uskml.zip"
if (!file.exists(fil)) download.file(url, fil)
unzip(fil, exdir="uskml")
# read KML file
us <- rgdal::readOGR("./uskml/cb_2014_us_nation_20m.kml", "cb_2014_us_nation_20m")
sp::proj4string(quakes) <- sp::proj4string(us)

length(quakes)
# 1514

usquakes = quakes[us,]
length(usquakes)
# 1260

### visualize
plot(us) 
# plot all quakes
points(quakes$longitude, quakes$latitude)
# plot just US
points(usquakes$longitude, usquakes$latitude)

结果图片：

谢谢@hrbrmstr！

Answer 1

这样做的方式是：

library(sp)
library(maptools)
library(rgeos) # not entirely necessary
library(rgdal) # not entirely necessary

url <- "http://earthquake.usgs.gov/earthquakes/feed/v1.0/summary/all_week.csv"
fil <- "all_week.csv"
if (!file.exists(fil)) download.file(url, fil)

quakes <- read.csv("all_week.csv", stringsAsFactors=FALSE)
coordinates(quakes) <- ~longitude+latitude

url <- "http://www2.census.gov/geo/tiger/GENZ2014/shp/cb_2014_us_nation_20m.zip"
fil <- "us.zip"
if (!file.exists(fil)) download.file(url, fil)
unzip(fil, exdir="us")
us <- readShapePoly("us/cb_2014_us_nation_20m.shp")

# alternatively
# us <- rgdal::readOGR("us/cbcb_2014_us_nation_20m.shp", "cb_2014_us_nation_20m")

# TRUE if in
in_us_rgeos <- rgeos::gIntersects(quakes, us, byid=TRUE)

# <NA> if in
in_us_over <- quakes %over% us

gIntersects需要更长时间。 rgdal和rgeos 可以成为在某些系统上工作的熊。 YMMV

使用美国KML需要（大部分）rgdal：

# you wanted KML shapefile tho
url <- "http://www2.census.gov/geo/tiger/GENZ2014/kml/cb_2014_us_nation_20m.zip"
fil <- "uskml.zip"
if (!file.exists(fil)) download.file(url, fil)
unzip(fil, exdir="uskml")

us <- rgdal::readOGR("uskml/cb_2014_us_nation_20m.kml", "cb_2014_us_nation_20m")
proj4string(quakes) <- proj4string(us)
rgeos::gIntersects(quakes, us, byid=TRUE)
head(quakes %over% us)