Spatial Data Integration with sf

library(linkeR)
library(shiny)
library(leaflet)
library(DT)
library(sf)

Introduction

The linkeR package provides seamless integration with sf spatial objects, automatically extracting coordinates from geometry columns for use in leaflet maps. This vignette demonstrates how to work with spatial data in linked dashboards.

Key Features

• Automatic coordinate extraction: linkeR automatically extracts longitude/latitude from sf geometry columns
• Preserves geometry: The original geometry column is kept for advanced spatial operations
• Mixed data types: Link sf objects with regular data frames
• Multi-geometry support: Works with POINT, POLYGON, and LINESTRING geometries

Basic SF Integration

Here’s a simple example using sf point data:

library(shiny)
library(leaflet)
library(DT)
library(linkeR)
library(sf)

# Create sample sf point data
create_sample_sf_data <- function() {
  # Create point geometries
  points <- st_sfc(
    st_point(c(-111.89, 40.76)),  # Salt Lake City
    st_point(c(-111.97, 41.22)),  # Ogden
    st_point(c(-111.66, 40.23)),  # Provo
    crs = 4326  # WGS84
  )
  
  # Create sf object
  sf_data <- st_sf(
    id = c("LOC_001", "LOC_002", "LOC_003"),
    name = c("Salt Lake City", "Ogden", "Provo"),
    population = c(200000, 87000, 116000),
    geometry = points
  )
  
  return(sf_data)
}

# Create corresponding table data
create_table_data <- function() {
  data.frame(
    id = c("LOC_001", "LOC_002", "LOC_003"),
    name = c("Salt Lake City", "Ogden", "Provo"),
    county = c("Salt Lake", "Weber", "Utah"),
    established = c(1847, 1851, 1849),
    area_sq_mi = c(111.1, 26.6, 44.2)
  )
}

ui <- fluidPage(
  titlePanel("SF Integration Demo"),
  
  fluidRow(
    column(6,
      h4("Spatial Data (SF Object)"),
      leafletOutput("location_map")
    ),
    column(6,
      h4("City Information"),
      DTOutput("city_table")
    )
  ),
  
  fluidRow(
    column(12,
      h4("Selection Details"),
      verbatimTextOutput("selection_info")
    )
  )
)

server <- function(input, output, session) {
  
  # Create reactive sf data
  sf_data <- reactive({
    create_sample_sf_data()
  })
  
  # Create reactive table data
  table_data <- reactive({
    create_table_data()
  })
  
  # Render leaflet map
  output$location_map <- renderLeaflet({
    data <- sf_data()
    
    # For initial rendering, extract coordinates manually
    coords <- st_coordinates(data)
    
    leaflet() %>%
      addTiles() %>%
      addCircleMarkers(
        lng = coords[, 1],
        lat = coords[, 2],
        layerId = data$id,  # Critical for linking!
        radius = 8,
        popup = ~paste("City:", data$name)
      ) %>%
      fitBounds(
        lng1 = min(coords[, 1]) - 0.1,
        lat1 = min(coords[, 2]) - 0.1,
        lng2 = max(coords[, 1]) + 0.1,
        lat2 = max(coords[, 2]) + 0.1
      )
  })
  
  # Render data table
  output$city_table <- renderDT({
    datatable(
      table_data(),
      selection = "single",
      rownames = FALSE,
      options = list(pageLength = 5)
    )
  })
  
  # Link sf object with regular data frame
  registry <- link_plots(
    session,
    location_map = sf_data,    # SF object - coordinates auto-extracted!
    city_table = table_data,   # Regular data frame
    shared_id_column = "id"
  )
  
  # Display selection information
  output$selection_info <- renderText({
    selection <- registry$get_selection()
    if (!is.null(selection$selected_id)) {
      selected_sf <- sf_data()[sf_data()$id == selection$selected_id, ]
      selected_table <- table_data()[table_data()$id == selection$selected_id, ]
      
      paste0(
        "Selected: ", selected_sf$name, "\n",
        "Source: ", selection$source, "\n",
        "Population: ", format(selected_sf$population, big.mark = ","), "\n",
        "County: ", selected_table$county, "\n",
        "Established: ", selected_table$established
      )
    } else {
      "No selection"
    }
  })
}

shinyApp(ui, server)

Advanced SF Integration with Custom Handlers

You can create custom click handlers that work with both the extracted coordinates and the original geometry:

# Custom click handler that uses both coordinates and geometry
custom_sf_handler <- function(map_proxy, selected_data, session) {
  if (!is.null(selected_data)) {
    # Use extracted coordinates for map operations
    longitude <- selected_data$longitude
    latitude <- selected_data$latitude
    
    # Create rich popup content
    popup_content <- paste0(
      "<div style='min-width: 200px;'>",
      "<h4>", selected_data$name, "</h4>",
      "<p><strong>Population:</strong> ", format(selected_data$population, big.mark = ","), "</p>",
      "<p><strong>Coordinates:</strong> ", round(longitude, 4), ", ", round(latitude, 4), "</p>",
      "<p><em>Data from SF object</em></p>",
      "</div>"
    )
    
    # Update map view and add popup
    map_proxy %>%
      leaflet::setView(lng = longitude, lat = latitude, zoom = 12) %>%
      leaflet::clearPopups() %>%
      leaflet::addPopups(
        lng = longitude,
        lat = latitude,
        popup = popup_content
      )
      
    # If you have the original geometry, you could also add buffers, etc.
    # This demonstrates how you can access both coordinate and geometry data
    
  } else {
    # Handle deselection
    map_proxy %>% leaflet::clearPopups()
  }
}

# Use the custom handler in link_plots
registry <- link_plots(
  session,
  location_map = sf_data,
  city_table = table_data,
  shared_id_column = "id",
  leaflet_click_handler = custom_sf_handler
)

Working with Different Geometry Types

linkeR supports various geometry types by extracting appropriate coordinates:

# Example with polygon data
create_polygon_sf_data <- function() {
  # Create polygon geometries (city boundaries)
  polygon1 <- st_polygon(list(cbind(
    c(-111.95, -111.85, -111.85, -111.95, -111.95),
    c(40.72, 40.72, 40.80, 40.80, 40.72)
  )))
  
  polygon2 <- st_polygon(list(cbind(
    c(-112.05, -111.95, -111.95, -112.05, -112.05),
    c(41.18, 41.18, 41.26, 41.26, 41.18)
  )))
  
  polygons <- st_sfc(polygon1, polygon2, crs = 4326)
  
  # Create sf object with polygon geometries
  sf_polygons <- st_sf(
    id = c("ZONE_001", "ZONE_002"),
    name = c("Downtown SLC", "Downtown Ogden"),
    area_sq_km = c(25.5, 18.2),
    geometry = polygons
  )
  
  return(sf_polygons)
}

# linkeR automatically extracts centroids for non-POINT geometries
polygon_data <- reactive({
  create_polygon_sf_data()
})

# This works the same way - linkeR handles the geometry type automatically
registry <- link_plots(
  session,
  zone_map = polygon_data,     # Polygon sf object
  zone_table = table_data,     # Regular data frame
  shared_id_column = "id"
)

Benefits of SF Integration

1. Automatic Coordinate Extraction

• No need to manually extract coordinates from geometry columns
• Works with POINT, POLYGON, and LINESTRING geometries
• Handles coordinate reference system transformations

2. Preserved Geometry

• Original geometry column is maintained
• Enables advanced spatial operations in custom handlers
• Supports complex spatial workflows

3. Mixed Data Types

• Link sf objects with regular data frames seamlessly
• Consistent behavior across different data sources
• Flexible dashboard design

4. Developer-Friendly

• Simple API - just pass your sf object to link_plots()
• Extensive error handling and validation
• Clear documentation and examples

Troubleshooting

Common Issues:

“Required columns missing”: Ensure your sf object has valid geometries that can be converted to coordinates.

“Non-POINT geometries detected”: This is normal - linkeR automatically uses centroids for polygons and lines. There is potential for more advanced support in the future.

Performance issues: Large sf objects can impact performance. Consider simplifying geometries or using spatial indexing.

Debugging SF Integration:

# Check your sf object structure
sf_data <- your_sf_object()
print(st_geometry_type(sf_data))  # Check geometry types
print(st_crs(sf_data))           # Check coordinate reference system
print(names(sf_data))            # Check column names

# Test coordinate extraction
coords <- st_coordinates(sf_data)
print(head(coords))              # Verify coordinates are numeric

Next Steps

• See vignette("getting-started") for basic linkeR usage
• See vignette("custom-behaviors") for advanced click handlers
• See vignette("multiple-components") for complex dashboard examples
• Run system.file("examples", "sf_app.R", package = "linkeR") for a complete sf demonstration

The sf integration in linkeR makes it easy to create interactive spatial dashboards without manual coordinate management, while preserving the full power of sf for advanced spatial operations.