Coordinate Reference Systems and Projections
Transform between CRS using st_crs() and st_transform() correctly.
Coordinate Reference Systems and Projections is a free R Academy lesson on CoddyKit. This is lesson 2 of 4. Any 3 lessons of this learning path are free to read in full — after that, CoddyKit PRO unlocks every lesson, plus hands-on practice with a built-in code editor and a 24/7 AI tutor. It is part of the R Academy learning path, and your progress syncs across the web and the CoddyKit app. The R Academy course includes 4 lessons in total.
Why CRS Matters
A Coordinate Reference System (CRS) tells R what the numbers in your geometry column mean — degrees of latitude/longitude, metres in a projected grid, or something else. Mixing data in different CRS without re-projecting causes spatial joins to fail silently and distance calculations to be wrong.
library(sf)
# Two sf objects in the same CRS: operations work
nc <- read_sf(system.file('shape/nc.shp', package = 'sf'))
# sf will error or warn when CRS do not match during operations
# This is the CRS metadata attached to the object
crs <- st_crs(nc)
cat('EPSG:', crs$epsg, '\n')
cat('Input is geographic (lon/lat)?', st_is_longlat(nc), '\n')
cat('Units:', crs$units_gdal, '\n')st_crs: Reading CRS Info
st_crs(sf_obj) returns a crs object with EPSG code, WKT string, Proj4 string, and unit information. You can also pass an integer EPSG code directly to retrieve the CRS definition.
library(sf)
nc <- read_sf(system.file('shape/nc.shp', package = 'sf'))
# Inspect the CRS object
crs <- st_crs(nc)
cat('Class:', class(crs), '\n')
cat('EPSG:', crs$epsg, '\n')
cat('Proj4:', substr(crs$proj4string, 1, 60), '...\n')
# Lookup a CRS by EPSG code
wgs84 <- st_crs(4326)
web_mercator <- st_crs(3857)
cat('\nWGS84 name:', wgs84$Name, '\n')
cat('Web Mercator name:', web_mercator$Name, '\n')EPSG:4326 — WGS84
EPSG:4326 is the World Geodetic System 1984 — the standard for GPS coordinates. Coordinates are in decimal degrees: longitude (X) ranges −180 to 180, latitude (Y) ranges −90 to 90. It is a geographic CRS, not projected.
library(sf)
# Create points in WGS84 (GPS coordinates)
cities <- data.frame(
city = c('London', 'Tokyo', 'Sydney', 'New York'),
lon = c(-0.118, 139.692, 151.209, -74.006),
lat = c(51.509, 35.689, -33.869, 40.713)
)
cities_sf <- st_as_sf(cities,
coords = c('lon', 'lat'),
crs = 4326
)
cat('CRS:', st_crs(cities_sf)$Name, '\n')
cat('Units:', st_crs(cities_sf)$units_gdal, '\n')
# Naive distance in degrees (not meaningful for real distances)
d <- st_distance(cities_sf[1, ], cities_sf[2, ])
cat('London-Tokyo distance (degrees):', round(as.numeric(d), 2), '\n')EPSG:3857 — Web Mercator
EPSG:3857 is the Web Mercator projection used by Google Maps, OpenStreetMap, and Leaflet. Coordinates are in metres, centred at the prime meridian / equator. It distorts areas near the poles but is good for web visualisation.
library(sf)
cities <- data.frame(
city = c('London', 'Tokyo'),
lon = c(-0.118, 139.692),
lat = c(51.509, 35.689)
)
cities_sf <- st_as_sf(cities, coords = c('lon', 'lat'), crs = 4326)
# Transform to Web Mercator
cities_merc <- st_transform(cities_sf, 3857)
cat('WGS84 coords (degrees):\n')
print(st_coordinates(cities_sf))
cat('\nWeb Mercator coords (metres):\n')
print(st_coordinates(cities_merc))
# Distance is now in metres
d <- st_distance(cities_merc[1, ], cities_merc[2, ])
cat('\nLondon-Tokyo distance (km):', round(as.numeric(d) / 1000, 0), '\n')st_set_crs: Assigning a CRS
st_set_crs(sf_obj, crs) assigns a CRS without transforming coordinates. Use this only when the CRS metadata is missing or wrong. If you want to re-project coordinates to a new CRS, use st_transform() instead.
library(sf)
# Create geometry with NO CRS
pts <- st_sfc(
st_point(c(-74.006, 40.713)),
st_point(c(-87.629, 41.878))
)
cat('CRS before:', is.na(st_crs(pts)), '(NA = no CRS)\n')
# Assign WGS84 (no coordinate transformation)
pts_wgs84 <- st_set_crs(pts, 4326)
cat('CRS after assignment:', st_crs(pts_wgs84)$epsg, '\n')
# Compare with st_transform (DOES transform coordinates)
pts_sf <- st_sf(id = 1:2, geometry = pts_wgs84)
pts_merc <- st_transform(pts_sf, 3857)
cat('Original lon:', st_coordinates(pts_sf)[1, 'X'], '\n')
cat('Mercator X: ', st_coordinates(pts_merc)[1, 'X'], '\n')st_transform: Re-Projecting
st_transform(sf_obj, crs) re-projects all geometries from their current CRS to the target CRS, applying the full geodetic transformation. The crs argument accepts an EPSG integer, a Proj4 string, or a WKT string.
library(sf)
nc <- read_sf(system.file('shape/nc.shp', package = 'sf'))
cat('Original CRS:', st_crs(nc)$epsg, '\n')
# Re-project to WGS84
nc_wgs84 <- st_transform(nc, 4326)
cat('Transformed CRS:', st_crs(nc_wgs84)$epsg, '\n')
# Re-project to a local projected CRS (NC State Plane)
nc_stateplane <- st_transform(nc, 32119) # NAD83 / North Carolina
cat('State Plane CRS:', st_crs(nc_stateplane)$Name, '\n')
# Area calculation is more accurate in a projected CRS
area_deg <- st_area(nc[1, ])
area_m2 <- st_area(nc_stateplane[1, ])
cat('Area (degrees^2):', format(area_deg, big.mark = ','), '\n')
cat('Area (m^2): ', format(area_m2, big.mark = ','), '\n')Choosing the Right Projection
Different projections optimise for different properties: equal-area (e.g. Mollweide), conformal (Mercator), equidistant. For local analysis use a projected CRS in metres; for global overviews WGS84 works; for web maps use Web Mercator.
library(sf)
# Common EPSG codes cheat sheet
epsg_table <- data.frame(
EPSG = c(4326, 3857, 4269, 32637, 27700),
Name = c(
'WGS84 (GPS, global)',
'Web Mercator (web maps)',
'NAD83 (North America geographic)',
'UTM zone 37N (Middle East)',
'British National Grid (UK)'
),
Units = c('degrees', 'metres', 'degrees', 'metres', 'metres')
)
print(epsg_table, row.names = FALSE)
# Look up any EPSG online: https://epsg.io/<code>
cat('\nTip: use st_crs(epsg_code)$Name to verify a code\n')Proj4 Strings
Before EPSG codes became standard, CRS were defined by Proj4 strings — compact parameter strings passed to the PROJ library. You can still use them with st_transform(), but EPSG codes are preferred for clarity and future compatibility.
library(sf)
nc <- read_sf(system.file('shape/nc.shp', package = 'sf'))
# WGS84 as Proj4 string
wgs84_proj4 <- '+proj=longlat +datum=WGS84 +no_defs'
nc_wgs84 <- st_transform(nc, crs = wgs84_proj4)
cat('Using Proj4 string:\n', wgs84_proj4, '\n')
cat('Result CRS EPSG:', st_crs(nc_wgs84)$epsg, '\n')
# Compare: equivalent to st_transform(nc, 4326)
cat('Same as EPSG 4326?',
identical(st_crs(nc_wgs84), st_crs(st_transform(nc, 4326))), '\n')Checking CRS Equality
Before any spatial join or overlay operation, verify that both objects share the same CRS using st_crs(a) == st_crs(b). If they differ, transform one to match the other before proceeding.
library(sf)
# Two objects in different CRS
nc_nad27 <- read_sf(system.file('shape/nc.shp', package = 'sf'))
nc_wgs84 <- st_transform(nc_nad27, 4326)
cat('Same CRS?', st_crs(nc_nad27) == st_crs(nc_wgs84), '\n')
# Always check before join/overlay
align_crs <- function(a, b) {
if (st_crs(a) != st_crs(b)) {
message('CRS mismatch! Transforming b to match a.')
b <- st_transform(b, st_crs(a))
}
list(a = a, b = b)
}
aligned <- align_crs(nc_nad27, nc_wgs84)
cat('After alignment - same CRS?',
st_crs(aligned$a) == st_crs(aligned$b), '\n')Calculating Distances with CRS
Distances computed on geographic CRS (degrees) are meaningless. Re-project to an equal-distance or equal-area projected CRS first, or use sf's built-in geodesic distance calculation which handles WGS84 correctly via the PROJ library.
library(sf)
# City points in WGS84
cities <- data.frame(
city = c('Paris', 'Berlin', 'Madrid', 'Rome'),
lon = c(2.350, 13.404, -3.702, 12.496),
lat = c(48.865, 52.520, 40.417, 41.902)
)
cities_sf <- st_as_sf(cities, coords = c('lon', 'lat'), crs = 4326)
# sf computes geodesic distances on WGS84 automatically
dist_matrix <- st_distance(cities_sf)
rownames(dist_matrix) <- cities$city
colnames(dist_matrix) <- cities$city
# Convert to km
dist_km <- round(as.numeric(dist_matrix) / 1000)
mat_km <- matrix(dist_km, 4, 4, dimnames = list(cities$city, cities$city))
print(mat_km)A Complete CRS Workflow
Here is a complete workflow: load raw GPS data, assign WGS84, re-project to a local metric CRS for accurate area/distance calculations, perform the analysis, and re-project back to WGS84 for web mapping output.
library(sf)
# Load and assign CRS
nc <- read_sf(system.file('shape/nc.shp', package = 'sf'))
nc_wgs84 <- st_transform(nc, 4326)
# Re-project to NC State Plane (metres) for accurate areas
nc_proj <- st_transform(nc, 32119)
# Compute area per county in km2
nc_proj$area_km2 <- as.numeric(st_area(nc_proj)) / 1e6
# Top 5 largest counties
top5 <- nc_proj[order(-nc_proj$area_km2), c('NAME', 'area_km2')]
cat('5 Largest NC Counties (km2):\n')
print(head(as.data.frame(top5)[, c('NAME', 'area_km2')], 5))
# Back to WGS84 for output
nc_output <- st_transform(nc_proj, 4326)
cat('\nOutput CRS:', st_crs(nc_output)$epsg, '\n')Quick Check
You need to compute accurate areas (in square metres) for polygon features currently in WGS84 (EPSG:4326). What should you do?
Recap: CRS and Projections
Key takeaways:
st_crs(sf_obj)reads CRS metadata; EPSG codes are the preferred way to specify CRS- EPSG:4326 = WGS84 (degrees, for GPS); EPSG:3857 = Web Mercator (metres, for web maps)
st_set_crs()assigns CRS without transforming;st_transform()re-projects coordinates- Always verify CRS equality with
st_crs(a) == st_crs(b)before spatial operations - For accurate distances and areas, re-project to a metric projected CRS first
- Proj4 strings still work but EPSG codes are cleaner and more readable
library(sf)
nc <- read_sf(system.file('shape/nc.shp', package = 'sf'))
# CRS pipeline
cat('Original:', st_crs(nc)$epsg, '\n')
nc_wgs84 <- st_transform(nc, 4326)
nc_proj <- st_transform(nc_wgs84, 32119)
cat('WGS84:', st_crs(nc_wgs84)$epsg, '\n')
cat('Projected:', st_crs(nc_proj)$epsg, '\n')
Learn R with an AI tutor — free
Write and run real code in your browser, get instant help from a 24/7 AI tutor, and pick up where you left off on the web or in the app.
- Courses
- 43
- Lessons
- 159
Frequently Asked Questions
Is the “Coordinate Reference Systems and Projections” lesson free?
Yes — any 3 lessons of the R Academy learning path, including “Coordinate Reference Systems and Projections”, are free to read in full here on the web. After that, CoddyKit PRO unlocks every lesson, plus interactive practice with a built-in code editor and a 24/7 AI tutor. The R Academy course includes 4 lessons in total.
What will I learn in “Coordinate Reference Systems and Projections”?
Transform between CRS using st_crs() and st_transform() correctly. You practise R Academy with hands-on code you run directly in the browser, and a 24/7 AI tutor answers your questions as you work through the lesson.
Do I need any experience to start R Academy?
No prior experience is required. R Academy on CoddyKit is structured for beginners through advanced learners, so you can start here or from the beginning and move at your own pace. This is lesson 2 of 4.
How long does the “Coordinate Reference Systems and Projections” lesson take?
Most CoddyKit lessons take about 5–10 minutes. Each one is bite-sized and interactive, so you make steady progress and pick up exactly where you left off across the web and the app.
Can I write and run code in this R Academy lesson?
Yes. Every R Academy lesson includes a built-in code editor, so you write and run real code right in your browser and get instant AI feedback — no local setup required.
All lessons in this course
- Simple Features with the sf Package
- Coordinate Reference Systems and Projections
- Spatial Joins and Operations
- Interactive Maps with leaflet