| name | sf-terra |
| description | R spatial data: sf vectors, terra rasters, spdep/spatialreg spatial stats, leaflet interactive maps, ggplot2+geom_sf() choropleths. CRS, spatial joins, geometry ops. Use when execution language is R. Python equivalent: geopandas. |
| autoload | never |
| metadata | {"audience":"research-coders","domain":"r-library","library-version":"sf 1.1-0","skill-last-updated":"2026-07-07","tags":["r","geospatial","sf","terra","mapping"]} |
sf-terra Skill
R geospatial analysis with sf (Simple Features for R), terra (modern raster
operations), spdep/spatialreg (spatial statistics and spatial econometrics),
leaflet (interactive web maps), and classInt (classification breaks for
choropleth maps). Covers sf objects, CRS management, spatial joins, geometric
operations, raster creation and manipulation, choropleth maps via
ggplot2 + geom_sf(), interactive maps via leaflet, and spatial autocorrelation
and regression. Use when the execution language is R and the task involves
spatial data, map-making, or spatial statistics. Python equivalent: the
geopandas skill.
What is sf?
sf (Simple Features) is the modern R package for vector geospatial data:
- sf objects: An sf object is a data.frame (or tibble) with a special
geometry list-column containing sfc geometries -- tabular data meets spatial operations
- Tidyverse integration: sf objects work with dplyr verbs (
filter(), mutate(), group_by(), summarize()), and the geometry column is preserved automatically
- ggplot2 integration:
geom_sf() plots sf objects with automatic CRS-aware axis labels
- GDAL/GEOS/PROJ backend: Full access to industry-standard spatial libraries
- Successor to sp/rgdal/rgeos: Those packages are retired; sf is the standard
What is terra?
terra is the modern R package for raster data (replaces the retired raster package):
- SpatRaster objects: Multi-band, multi-layer raster data with CRS and extent metadata
- Raster arithmetic: Element-wise operations, aggregation, focal operations
- Raster-vector interaction:
extract(), crop(), mask() using sf geometries
- Memory-efficient: Processes large rasters on disk without loading everything into memory
Version Notes
Versions installed in the DAAF container (R 4.5.3):
| Package | Version | Role |
|---|
| sf | 1.1-0 | Vector spatial data (Simple Features) |
| terra | 1.9-11 | Raster spatial data |
| stars | 0.7.2 | Spatiotemporal arrays (raster time series) |
| spdep | 1.4.2 | Spatial weights and spatial autocorrelation |
| spatialreg | 1.4.3 | Spatial regression models (SAR, SEM, SDM) |
| leaflet | 2.2.3 | Interactive web maps |
| classInt | 0.4.11 | Classification intervals for choropleth breaks |
sf 1.x changes to be aware of:
- s2 geometry engine enabled by default for geographic CRS (spherical geometry) -- this changes area/distance/intersection behavior from earlier versions
- Use
sf_use_s2(FALSE) to temporarily revert to planar GEOS operations when s2 produces unexpected results (see gotchas.md)
st_make_valid() is now built into sf (no longer needs lwgeom)
Also Available in This Environment
Four additional spatial packages are installed (versions verified) that map to
capabilities the geopandas skill documents on the Python side:
| Package | Version | Purpose | geopandas-skill equivalent |
|---|
| maptiles | 0.11.0 | Basemap tiles for static maps (get_tiles()) | contextily |
| exactextractr | 0.10.1 | Fast, exact zonal statistics of rasters over polygons (exact_extract()) | rasterstats zonal stats |
| osmdata | 0.3.0 | Download OpenStreetMap features via the Overpass API | osmnx / OSM download |
| tidygeocoder | 1.0.6 | Geocoding addresses to coordinates (geocode()) | geopy geocoding |
Honest asymmetry: point-pattern analysis (Ripley's K/G/F functions, spatial
point process models) requires spatstat, which is not installed — for
point-pattern methods, Python (PySAL pointpats via the geopandas skill) is
the routed path. Areal interpolation (tobler-style) also has no installed R
equivalent.
How to Use This Skill
Reference File Structure
| File | Purpose | When to Read |
|---|
quickstart.md | sf objects, st_read/st_write, basic plotting, CRS inspection | Starting with sf or quick reminder |
crs.md | st_transform(), EPSG codes, CRS management, projection pitfalls | CRS errors or projection decisions |
spatial-ops.md | st_join(), st_buffer(), st_intersection(), st_union(), geometric operations | Combining or transforming spatial data |
raster.md | terra::rast(), raster arithmetic, extract(), crop/mask, raster-vector interaction | Working with raster data |
mapping.md | ggplot2 + geom_sf(), choropleth with scale_fill, classInt for breaks, annotation | Making static maps |
interactive.md | leaflet: addPolygons, addCircleMarkers, colorBin, popups, tile layers | Making interactive web maps |
spatial-stats.md | spdep::poly2nb(), nb2listw(), moran.test(), spatialreg::lagsarlm() | Spatial autocorrelation and regression |
gotchas.md | s2 geometry, CRS mismatch, geometry column persistence, large data performance | Debugging spatial issues |
Reading Order
- New to sf? Start with
quickstart.md then spatial-ops.md
- Making maps? Read
mapping.md (relies on crs.md for projection choices)
- Interactive maps? Read
interactive.md
- Raster data? Read
raster.md
- Spatial statistics? Read
spatial-stats.md (for methodology context, also load data-scientist skill's geospatial-analysis.md)
- Having issues? Check
gotchas.md first
Related Skills
| Skill | Relationship |
|---|
geopandas | Python equivalent -- covers the same spatial domain for Python pipelines |
data-scientist (geospatial-analysis.md, geospatial-operations.md) | Spatial methodology -- when/why to use spatial methods, interpretation guidance |
ggplot2 | Visualization -- geom_sf() is built on ggplot2; load ggplot2 skill for themes and styling |
tidyverse | Data preparation -- sf objects work with dplyr verbs; tidyverse prepares data for spatial analysis |
r-stats | Statistical modeling -- non-spatial regression and diagnostics |
fixest | Fixed effects regression -- when combining spatial data with panel econometrics |
r-python-translation | Cross-language spatial translation |
Quick Decision Trees
"I need to read or write spatial data"
Loading/saving spatial data?
+-- Read vector file (Shapefile, GeoPackage, GeoJSON) -> ./references/quickstart.md
+-- Read GeoParquet -> ./references/quickstart.md
+-- Create sf from lat/lon columns -> ./references/quickstart.md
+-- Write to file -> ./references/quickstart.md
+-- Read raster data (GeoTIFF) -> ./references/raster.md
+-- Raster time series (NetCDF) -> ./references/raster.md
"I need to combine or transform spatial data"
Spatial operations?
+-- Join by location (point-in-polygon) -> ./references/spatial-ops.md
+-- Nearest-neighbor join -> ./references/spatial-ops.md
+-- Overlay (intersection, union, difference) -> ./references/spatial-ops.md
+-- Dissolve/aggregate polygons -> ./references/spatial-ops.md
+-- Buffer features -> ./references/spatial-ops.md
+-- Clip to boundary -> ./references/spatial-ops.md
+-- Compute distances -> ./references/spatial-ops.md
+-- Compute centroids or areas -> ./references/spatial-ops.md
"I need to fix CRS or projection issues"
CRS/projection issues?
+-- Check current CRS -> ./references/crs.md
+-- Reproject to different CRS -> ./references/crs.md
+-- Choose a projection for analysis -> ./references/crs.md
+-- Data has no CRS (set it) -> ./references/crs.md
+-- CRS mismatch error -> ./references/gotchas.md
+-- Area/distance calculations wrong -> ./references/crs.md
"I need to make a map"
Making maps?
+-- Quick static choropleth (ggplot2) -> ./references/mapping.md
+-- Classification schemes (quantiles, Jenks) -> ./references/mapping.md
+-- Layered map with multiple sf layers -> ./references/mapping.md
+-- Multi-panel maps -> ./references/mapping.md
+-- Interactive map (pan/zoom/hover) -> ./references/interactive.md
+-- Export to PNG/SVG/HTML -> ./references/mapping.md or ./references/interactive.md
"I need spatial statistics"
Spatial statistics?
+-- Build spatial weights (contiguity, distance) -> ./references/spatial-stats.md
+-- Test for spatial autocorrelation (Moran's I) -> ./references/spatial-stats.md
+-- Local clusters / hot spots (LISA) -> ./references/spatial-stats.md
+-- Spatial regression (lag, error, Durbin) -> ./references/spatial-stats.md
+-- Methodology guidance (interpretation, MAUP) -> data-scientist skill: geospatial-analysis.md
"I need to work with rasters"
Raster operations?
+-- Read GeoTIFF -> ./references/raster.md
+-- Raster arithmetic -> ./references/raster.md
+-- Extract raster values at points/polygons -> ./references/raster.md
+-- Crop/mask raster by polygon -> ./references/raster.md
+-- Raster-vector conversion -> ./references/raster.md
+-- Raster time series (stars) -> ./references/raster.md
"Something isn't working"
Having issues?
+-- s2 geometry errors -> ./references/gotchas.md
+-- CRS mismatch errors -> ./references/gotchas.md
+-- Geometry column lost after dplyr -> ./references/gotchas.md
+-- Invalid geometry errors -> ./references/gotchas.md
+-- Memory issues with large files -> ./references/gotchas.md
+-- General troubleshooting -> ./references/gotchas.md
File-First Execution in Research Workflows
In DAAF research pipelines, R spatial operations follow the file-first execution
protocol -- code is written to .R script files and executed via the
run_with_capture.sh wrapper, never run interactively.
The pattern:
- Write spatial analysis code to
scripts/stage{N}_{type}/{step}_{task-name}.R
- Execute via Bash:
bash {BASE_DIR}/scripts/run_with_capture.sh {PROJECT_DIR}/scripts/{script_name}.R
run_with_capture.sh detects the .R extension and uses Rscript automatically- stdout/stderr are appended to the script file as comments
- If a script fails, create a versioned copy (
_a.R, _b.R, etc.) for fixes
Read agent_reference/SCRIPT_EXECUTION_REFERENCE.md for the complete protocol.
R spatial script structure follows DAAF conventions:
library(sf)
library(dplyr)
library(ggplot2)
PROJECT_DIR <- "/daaf/research/YYYY-MM-DD_Project"
TARGET_CRS <- 5070
schools <- st_read(file.path(PROJECT_DIR, "data", "schools.gpkg")) |>
st_transform(TARGET_CRS)
counties <- st_read(file.path(PROJECT_DIR, "data", "counties.gpkg")) |>
st_transform(TARGET_CRS)
cat("Schools:", nrow(schools), "rows\n")
cat("Counties:", nrow(counties), "rows\n")
joined st_joinschools counties join st_within
stopifnotnrowjoined nrowschools
stopifnotjoinedcounty_name
cat nrowjoined
st_writejoined file.pathPROJECT_DIR
delete_dsn
cat
Quick Reference
Essential Setup
library(sf)
library(terra)
library(dplyr)
library(ggplot2)
library(spdep)
library(spatialreg)
library(leaflet)
library(classInt)
Core Operations
| Operation | Code | Package |
|---|
| Read vector | st_read("data.gpkg") | sf |
| Read parquet (WKB geometry) | arrow::read_parquet() + st_as_sfc(structure(as.list(col), class = "WKB"), crs = ...) — sfarrow NOT installed; see quickstart.md | sf + arrow |
| From lat/lon | st_as_sf(df, coords = c("lon", "lat"), crs = 4326) | sf |
| Check CRS | st_crs(x) | sf |
| Reproject | st_transform(x, crs = 5070) | sf |
| Spatial join | st_join(points, polygons, join = st_within) | sf |
| Intersection | st_intersection(x, y) | sf |
| Buffer | st_buffer(x, dist = 1000) (CRS units) | sf |
| Centroid | st_centroid(x) | sf |
| Area | st_area(x) (returns units) | sf |
| Distance | st_distance(x, y) | sf |
| Union/dissolve | `x | > group_by(col) |
| Plot (static) | ggplot(x) + geom_sf(aes(fill = col)) | ggplot2 |
| Plot (interactive) | `leaflet(x) | > addTiles() |
| Write vector | st_write(x, "out.gpkg") | sf |
| Read raster | rast("elev.tif") | terra |
| Raster extract | extract(r, vect(sf_obj)) | terra |
| Crop raster | crop(r, vect(sf_obj)) | terra |
Common CRS Codes
| EPSG | Name | Use For |
|---|
| 4326 | WGS84 | Storage, exchange, web (not analysis) |
| 5070 | NAD83 Conus Albers | US thematic maps (equal-area) |
| 3857 | Web Mercator | Web tiles display only |
| 32617 | UTM Zone 17N | US East Coast local analysis |
Topic Index
| Topic | Reference File |
|---|
| sf objects and structure | ./references/quickstart.md |
| st_read / st_write | ./references/quickstart.md |
| Creating sf from coordinates | ./references/quickstart.md |
| Basic plotting with plot() | ./references/quickstart.md |
| GeoParquet I/O | ./references/quickstart.md |
| CRS fundamentals | ./references/crs.md |
| st_transform / reprojection | ./references/crs.md |
| Choosing projections | ./references/crs.md |
| Common US projections | ./references/crs.md |
| set_crs vs st_transform | ./references/crs.md |
| Spatial joins (st_join) | ./references/spatial-ops.md |
| Nearest-neighbor joins | ./references/spatial-ops.md |
| Overlay (st_intersection, st_union) | ./references/spatial-ops.md |
| Dissolve and aggregation | ./references/spatial-ops.md |
| Buffering (st_buffer) | ./references/spatial-ops.md |
| Clipping (st_crop, st_intersection) | ./references/spatial-ops.md |
| Area, distance, centroid | ./references/spatial-ops.md |
| terra::rast() basics | ./references/raster.md |
| Raster arithmetic | ./references/raster.md |
| extract(), crop(), mask() | ./references/raster.md |
| stars spatiotemporal arrays | ./references/raster.md |
| Raster-vector conversion | ./references/raster.md |
| ggplot2 + geom_sf() maps |
Citation
When sf is used as a primary analytical tool, include in the report's
Software & Tools references:
Pebesma, E. (2018). "Simple Features for R: Standardized Support for Spatial Vector Data." The R Journal, 10(1), 439-446. https://doi.org/10.32614/RJ-2018-009
Cite when: sf is used for spatial operations, spatial joins, or map
production central to the analysis.
Do not cite when: Only used to read a shapefile for a simple reference lookup.
If terra raster operations are also used, additionally cite:
Hijmans, R.J. (2024). terra: Spatial Data Analysis. R package. https://rspatial.org/
If spdep/spatialreg spatial analysis is also used, additionally cite:
Bivand, R.S., Pebesma, E., and Gomez-Rubio, V. (2013). Applied Spatial Data Analysis with R (2nd ed.). Springer. https://doi.org/10.1007/978-1-4614-7618-4
