使用 geom_sf 将纬度/经度点转换为地图

Question

I have a data set of latitude/longitude points that seek to convert to a simple feature (sf) in R. My goal is to plot these locations on a US map with polygons retrieved from the urbnmapr library.

Plotting with our a geographic reference, as shown in the code, results in all points being displayed.

When the points are plotted using geom_sf() they end up in South Dakota. It seems the latitude/longitude points are not being converted into the correct coordinate reference system, despite what I think is the correct use of the st_as_sf() function.

What correction needs to be made to this code to show the distribution of wind turbine locations properly on the US map?

# Map the locations of US Wind Turbines
library(urbnmapr)
library(ggplot2)
library(readr)
library(dplyr)
library(sf)

# This file is available from https://eerscmap.usgs.gov/uswtdb/assets/data/uswtdbCSV.zip
turbine <- read_csv("C:\\mydir\\uswtdb_v3_1_20200717.csv")

# Convert lat/long to a sf
turbine_sf <- turbine %>%
  st_as_sf(coords = c("xlong", "ylat"), crs=2163)

# obtain state geometries
states_sf <- get_urbn_map(map = "states", sf = TRUE)

# Remove AK, HI from state and PR and GU from turbines as well
states_sf <- states_sf[!(states_sf$state_abbv %in% c("HI","AK")),]
turbine   <- turbine  [!(turbine$t_state      %in% c('HI','AK','PR','GU')),]

# simple plot shows all locations
ggplot(turbine, aes(x=xlong, y=ylat)) + geom_point()

#plot locations over map
  ggplot() +
  geom_sf(data = turbine_sf) + 
  geom_sf(data = states_sf, fill = NA, color = "black", size = 0.15, alpha = 0) +
  coord_sf(datum = st_crs(2163)) +   
  labs(fill  = "", 
       title = "",
       caption='') + 
  theme_bw()

Answer 1

Your turbine dataset contains "xlong" and "ylat" in degrees ie geographic coordinate system with WGS84 datum (EPSG code: 4326). So, first, make it as crs = 4326 and then use st_transform(turbine_sf, crs=2163) to make same coordinate system with states_sf . You can use the following code

# Map the locations of US Wind Turbines
library(urbnmapr)
library(ggplot2)
library(readr)
library(dplyr)
library(sf)

# This file is available from https://eerscmap.usgs.gov/uswtdb/assets/data/uswtdbCSV.zip
turbine <- read_csv("uswtdb_v3_1_20200717.csv")

# Convert lat/long to a sf
turbine_sf <- turbine %>%
  st_as_sf(coords = c("xlong", "ylat"), crs=4326)

turbine_sf_t <- st_transform(turbine_sf, crs=2163)
# obtain state geometries
states_sf <- get_urbn_map(map = "states", sf = TRUE)

st_crs(states_sf)
# Remove AK, HI from state and PR and GU from turbines as well
states_sf <- states_sf[!(states_sf$state_abbv %in% c("HI","AK")),]
turbine   <- turbine  [!(turbine$t_state      %in% c('HI','AK','PR','GU')),]

# simple plot shows all locations
ggplot(turbine, aes(x=xlong, y=ylat)) + geom_point()

#plot locations over map
ggplot() +
  geom_sf(data = turbine_sf_t) + 
  geom_sf(data = states_sf, fill = NA, color = "black", size = 0.15, alpha = 0) +
  coord_sf(datum = st_crs(2163)) +   
  labs(fill  = "", 
       title = "",
       caption='') + 
  theme_bw()

Answer 2

By doing st_as_sf(coords = c("xlong", "ylat"), crs=2163) you're saying that the original long, lat from your turbine table are based on CRS of 2163. I think you want to set them as 4326 which is the long lat under WGS84.

After setting the initial CRS, use st_transform() to transform the CRS of your shape to new CRS, eg turbine_sf <- st_transform(turbine_sf, crs=2163)