当我尝试按 ID 计算区域重叠时，group_by 函数不起作用答案

【问题标题】：The group_by function is not working when I try to calculate area overlap by ID当我尝试按 ID 计算区域重叠时，group_by 函数不起作用
【发布时间】：2020-06-12 10:18:43
【问题描述】：

我有 GPS 数据，可用于制作家庭范围。我想测量范围与保护区形状文件 (available here) 的重叠。当我有一个人时这很好用，但当我尝试使用多个 id 时就不行了。

问题似乎出在这段代码中的 group_by 函数上：

trk %>% group_by(id) %>% 
  hr_kde(., levels = c(.95)) %>% 
  hr_isopleths(.) %>% 
  st_intersection(., merged_Africa_tranform) %>% 
  st_area(.) / 1e6

它只会产生一个值（29.77905），而我希望有两个，每个 id 一个（A 为 34.68964，B 为 38.99062）。

这是完整的代码：

# load packages
library(tidyverse)
library(amt)
library(sf)
library(adehabitatHR)

#' load in the protected areas (working with Albers Equal Area)
merged_Africa = read_sf("shape//eswatini//WDPA_Apr2020_SWZ-shapefile-polygons.shp")
st_crs(merged_Africa) <- 4326
merged_Africa_tranform <- st_transform(merged_Africa, "ESRI:102022")

#' try a data frame with multiple IDs
x_ <- c(707692, 707589, 707998, 708407, 708916, 709415, 707743, 707429, 707971, 708143, 708981, 709156)
y_ <- c(-3030991,-3031423,-3031640,-3031750,-3032508,-3037158,-3030995,-3031723,-3031680,-3031755,-3032408,-3037758)
id <- c(rep("A",6), rep("B", 6))

mydata <- data.frame(x_,y_,id)

# transform to trk object
trk <-
  mk_track(mydata,
           .x = x_,
           .y = y_,
           id = id,
           crs = CRS("+proj=aea +lat_1=20 +lat_2=-23 +lat_0=0 +lon_0=25 +x_0=0 +y_0=0 +ellps=WGS84 +datum=WGS84 +units=m +no_defs"))

#' this should group by id, extract the home range, convert to polygon and measure the overlap
trk %>% group_by(id) %>% 
  hr_kde(., levels = c(.95)) %>% 
  hr_isopleths(.) %>% 
  st_intersection(., merged_Africa_tranform) %>% 
  st_area(.) / 1e6 #' only produces one value which is the same if id is ignored

【问题讨论】：

标签： r dplyr gis pipeline spatial

【解决方案1】：

我来回答你的问题..

# load packages
library(tidyverse)
library(amt)
library(sf)
library(adehabitatHR)

#loading and transforming shapefile can be simplified
merged_Africa <- read_sf("./temp/WDPA_Jun2020_SWZ-shapefile-polygons.shp") %>% 
  st_set_crs( 4326 ) %>% 
  st_transform( 102022 )

#mydata comes from sample data provided
#loop over the id's, crate track by ID, and calulate (and sum) overlapping area's
L <- lapply( unique( mydata$id ), function(x) {
  track_id <- mk_track( mydata[ id == x, ],
                        .x = x_,
                        .y = y_,
                        id = id,
                        crs = CRS("+proj=aea +lat_1=20 +lat_2=-23 +lat_0=0 +lon_0=25 +x_0=0 +y_0=0 +ellps=WGS84 +datum=WGS84 +units=m +no_defs"))
  track_object <- track_id %>% hr_kde(., levels = c(.95)) %>% hr_isopleths(.)
  area_totals <- track_object %>% st_intersection( merged_Africa ) %>% st_area()
  sum( area_totals )
  }) 
#set names
names(L) <- unique( mydata$id )

L

# $A
# 34689637 [m^2]
# 
# $B
# 38990619 [m^2]

【讨论】：