Vector Tile Server for Spatial Databases in R
rtileserver provides a lightweight HTTP server for serving Mapbox
Vector Tiles (MVT) from spatial databases including DuckDB and
PostgreSQL/PostGIS. The package enables efficient streaming of spatial
data to web mapping applications using the standard {z}/{x}/{y} tile
URL format.
# Install from GitHub
remotes::install_github("arthurgailes/rtileserver")
# Examples require the development version of mapgl:
# remotes::install_github("walkerke/mapgl")- Simple API: Start and stop tile servers with two functions
- Database agnostic: Works with any DBI-compatible spatial database
that supports
ST_AsMVT - Automatic port finding: Finds available ports automatically or use custom ports
- CORS enabled: Ready for web mapping applications
- Lightweight: Minimal dependencies using httpuv
This example uses the built-in North Carolina dataset from the sf
package - you can run this immediately!
library(DBI)
library(duckdb)
#> Warning: package 'duckdb' was built under R version 4.5.1
library(sf)
#> Warning: package 'sf' was built under R version 4.5.1
#> Linking to GEOS 3.13.1, GDAL 3.11.0, PROJ 9.6.0; sf_use_s2() is TRUE
library(duckspatial)
#> Warning: package 'duckspatial' was built under R version 4.5.1
# library(rtileserver)
devtools::load_all(".")
#> ℹ Loading rtileserver
# Load North Carolina counties data (built into sf package)
nc <- st_read(system.file("shape/nc.shp", package = "sf"))
#> Reading layer `nc' from data source
#> `C:\Users\arthur.gailes\AppData\Local\R\win-library\4.5\sf\shape\nc.shp'
#> using driver `ESRI Shapefile'
#> Simple feature collection with 100 features and 14 fields
#> Geometry type: MULTIPOLYGON
#> Dimension: XY
#> Bounding box: xmin: -84.32385 ymin: 33.88199 xmax: -75.45698 ymax: 36.58965
#> Geodetic CRS: NAD27
nc_merc <- st_transform(nc, 3857)
# Create DuckDB connection
con <- dbConnect(duckdb::duckdb(), dbdir = ":memory:")
# Install and load spatial extension using duckspatial
ddbs_install(con)
#> ℹ spatial extension version <a6a607f> is already installed in this database
ddbs_load(con)
#> ✔ Spatial extension loaded
# Write spatial data to DuckDB using duckspatial
ddbs_write_vector(con, nc_merc, "nc_counties", overwrite = TRUE)
#> ℹ Table nc_counties dropped
#> ✔ Table nc_counties successfully imported
# Optional: Create spatial index for better performance
dbExecute(con, "CREATE INDEX idx_nc_geom ON nc_counties USING RTREE (geometry)")
#> [1] 0
# Start tile server
server <- start_tile_server(
con,
table_name = "nc_counties",
geometry_column = "geometry",
properties = c("NAME", "FIPS", "AREA")
)
#> Tile server running at http://127.0.0.1:8001/
#> Tiles available at: http://127.0.0.1:8001/tiles/{z}/{x}/{y}.pbf
# Server info
print(server)
#> <rtileserver>
#> Server: http://127.0.0.1:8001/
#> Tiles: http://127.0.0.1:8001/tiles/{z}/{x}/{y}.pbf
#> Table: nc_counties
# Tiles are now available at server$url
# Use them in mapgl, Leaflet, Mapbox GL JS, etc.The tiles can be used with any mapping library that supports Mapbox Vector Tiles.
Continuing from the Quick Start example:
library(mapgl)
# Create interactive map
map <- maplibre(
center = c(-79.5, 35.5),
zoom = 6,
style = carto_style("positron")
) |>
add_vector_source(
id = "nc-tiles",
tiles = server$url,
minzoom = 0,
maxzoom = 14
) |>
add_fill_layer(
id = "nc-fill",
source = "nc-tiles",
source_layer = "layer",
fill_color = "steelblue",
fill_opacity = 0.6,
tooltip = "NAME"
) |>
add_line_layer(
id = "nc-outline",
source = "nc-tiles",
source_layer = "layer",
line_color = "white",
line_width = 1
)
# Display the map
map
When finished, disconnect the server:
# Clean up when done
stop_tile_server(server, disconnect_db = TRUE)library(RPostgres)
library(rtileserver)
# Connect to PostgreSQL
con <- dbConnect(
RPostgres::Postgres(),
dbname = "mydb",
host = "localhost",
user = "user",
password = "password"
)
# Start tile server
server <- start_tile_server(
con,
table_name = "my_spatial_table",
geometry_column = "geom", # Default is "geometry"
layer_name = "my_layer" # Default is "layer"
)
# Use tiles...
# Clean up
stop_tile_server(server, disconnect_db = TRUE)- R >= 4.0.0
- DBI package
- httpuv package
- A spatial database with
ST_AsMVTsupport:- DuckDB >= 1.4.0 with spatial extension
- PostgreSQL with PostGIS extension
- Spatial data should be in Web Mercator projection (EPSG:3857) for optimal performance
- Tables must have a geometry column (default name: “geometry”)
- Use spatial indexes for better performance
Start a vector tile server.
Arguments: - con: Database connection object (DBI-compatible) -
table_name: Name of the spatial table - geometry_column: Geometry
column name (default: “geometry”) - layer_name: MVT layer name
(default: “layer”) - properties: Character vector of columns to
include (default: NULL = all columns) - host: Host address (default:
“127.0.0.1”) - port: Port number (default: NULL = auto-detect)
Returns: An rtileserver object with server information
Stop a running tile server.
Arguments: - server: An rtileserver object - disconnect_db:
Whether to disconnect the database (default: FALSE)
Returns: NULL (invisibly)
-
Use spatial indexes: Create spatial indexes on your geometry columns
CREATE INDEX idx_geom ON features USING RTREE (geometry);
-
Project to Web Mercator: Transform data to EPSG:3857 before loading
data_mercator <- sf::st_transform(data, 3857)
-
Limit properties: Only include necessary columns in tiles
server <- start_tile_server(con, "features", properties = c("id", "name"))
-
Use appropriate zoom levels: Set minzoom/maxzoom based on data density
This package was inspired by Kyle
Walker’s demonstration of serving vector
tiles from DuckDB using the new ST_AsMVT function in DuckDB 1.4.0.
MIT
Contributions are welcome! Please feel free to submit issues or pull requests.