Spatial Data in R: Overview and Examples

Earth Observation Workshop, Basel

Michael Dorman
Geography and Environmental Development
dorman@post.bgu.ac.il

2018-02-26 - 2018-03-01
Contents

I The aim of this tutorial is to provide an overview to spatial

analysis in R
I Slides and code available on GitHub
I Data available on Dropbox
Requirements

I Several packages need to be installed to run code examples -

install.packages("rgdal")
install.packages("sf")
install.packages("raster")
install.packages("rgeos")
install.packages("geosphere")
install.packages("rmapshaper")
install.packages("gstat")
install.packages("automap")
install.packages("spdep")
install.packages("spatstat")
install.packages("osmdata")
install.packages("mapview")
Requirements

I Set to working directory with the data -

I cb_2015_us_state_5m - US states (Shapefile)
I us-states.geojson - US states (GeoJSON)
I dem.tif - Digital Elevation Model (GeoTIFF)
Introduction
I R is a programming language and environment, originally
developed for statistical computing and graphics
I Over time, there was an increasing number of contributed
packages for handling and analysing spatial data in R
I Today, spatial analysis is a major functionality in R
I As of November 2017, there are over 180 packages
specifically addressing spatial analysis in R

Figure 1: Books on Spatial Data Analysis with R

History

I pre-2003: Variable and incomplete approaches (MASS,

spatstat, maptools, geoR, splancs, gstat, . . . )
I 2003: Concensus that a package with base classes should be
useful; rgdal released on CRAN
I 2005: sp released on CRAN; sp support in rgdal
I 2008: Applied Spatial Data Analysis with R, 1st ed.
I 2010: raster released on CRAN
I 2011: rgeos released on CRAN
I 2013: Applied Spatial Data Analysis with R, 2nd ed.
I 2016: sf released on CRAN
History

Figure 2: The network structure of CRAN1 ; sp ecosystem shown in green

1
blog.revolutionanalytics.com/2015/07/the-network-structure-of-cran.html
R as a GIS?

I Strengths of R as a GIS
I R capabilities in data processing and visualization, combined
with dedicated packages for spatial data
I A single environment encompassing all analysis aspects -
acquiring data, computation, statistics, visualization, Web, etc.
I General advantages of programming
I Automation - Doing otherwise unfeasible repetitive tasks
I Reproducibility - Precise control of instructions to the
computer
I Situations when other tools are needed
I Interactive editing or georeferencing (but see mapedit)
I Unique GIS algorithms (label placement, network routing,
splitting lines at intersections, etc.)
I Data that cannot fit in RAM (but R can connect to spatial
databases)
Notable packages

I Handling spatial data

I sp, rgdal, rgeos, sf, raster
I Geometry
I geosphere, mapshaper
I Statistics
I gstat, automap, spatstat, spdep
I Data access
I osmdata
I Visualization
I ggplot2, ggmap, mapview, leaflet
Vector: Types of data

POINT LINESTRING POLYGON

MULTIPOINT MULTILINESTRING MULTIPOLYGON GEOMETRYCOLLECTION

Figure 3: Vector data types (Simple Features)

Vector: File formats

I Binary
I ESRI Shapefile (.shp, .shx, .dbf, .prj, . . . )
I Plain Text
I GeoJSON (.json or .geojson)
I GPX (.gpx)
I Spatial Databases
I PostGIS / PostgreSQL
Vector: Data structures (sp)

I Package sp has 6 main classes for vector layers

I One for each geometry type (points, lines, polygons)
I One for geometry only and one for geometry with attributes

Table 1: Spatial data structures in package sp

class geometry attributes

SpatialPoints Points -
SpatialPointsDataFrame Points data.frame
SpatialLines Lines -
SpatialLinesDataFrame Lines data.frame
SpatialPolygons Polygons -
SpatialPolygonsDataFrame Polygons data.frame
Vector: Data structures (sf)

I Package sf defines a hierarchical class system

I Class sfg is for a single geometry
I Class sfc is a set of geometries with a CRS
I Class sf is a layer with attributes

Table 2: Spatial data structures in package sf

class hierarchy data

sfg geometry coords, type, dimension
sfc geometry column Set of sfg + CRS
sf layer sfc + attributes
Raster: Types of data

Figure 4: Raster data types2

2
http://neondataskills.org/R/Introduction-to-Raster-Data-In-R/
Raster: File formats

I “Simple” rasters
I GeoTiff (.tif)
I Erdas Imagine Image (.img)
I “Complex” rasters (>3D and / or metadata)
I HDF (.hdf)
I NetCDF (.nc)
Raster: Data structures

I The raster package defines 3 classes for raster data

I RasterLayer for single-band
I RasterStack and RasterBrick for multi-band

Table 3: Spatial data structures in package raster

class layers storage

RasterLayer 1 Disk or RAM
RasterStack >=1 Disk and/or RAM
RasterBrick >=1 Disk (single file) or RAM
Input and output of spatial data

I Reading spatial layers from a file into an R data structure, or

writing the R data structure into a file, are handled by external
libraries -
I OGR is used for reading/writing vector files, with rgdal / sp
or sf
I GDAL is used for reading/writing raster files, with raster
I PROJ4 is used for handling CRS, in both rgdal / sp, sf and
raster
I Working with specialized formats, e.g. HDF with gdalUtils
or NetCDF with ncdf4
Coordinate Reference Systems (CRS)

I A Coordinate Reference System (CRS) defines how the

coordinates in the data relate to the surface of the Earth
I Two types of CRS
I Geographic - longitude and latitude, in degrees
I Projected - implying flat surface, associated with measurement
units (e.g. meters)
Coordinate Reference Systems (CRS)

Figure 5: Geographic (WGS 84; left) and projected (NAD83 / UTM zone
12N; right) CRS3

3
https://geocompr.robinlovelace.net/spatial-class.html
Coordinate Reference Systems (CRS)

I In R, CRS are defined using an ESPG code or PROJ.4 string

I Comparing Map Projections
I A table with EPSG code and PROJ.4 strings for all projections
-

crs = rgdal::make_EPSG()
Coordinate Reference Systems (CRS)

I For example, the WGS84 geographic projection can be

defined using its EPSG code

4326

I Or using its PROJ.4 string

"+proj=longlat +datum=WGS84 +no_defs"

rgdal, sf, raster: Handling spatial data

I Function readOGR can be used to read vector layers into sp

data structures -
library(rgdal)

pol_sp = readOGR("data", "cb_2015_us_state_5m")

## OGR data source with driver: ESRI Shapefile
## Source: "data", layer: "cb_2015_us_state_5m"
## with 49 features
## It has 1 fields

class(pol_sp)
## [1] "SpatialPolygonsDataFrame"
## attr(,"package")
## [1] "sp"
rgdal, sf, raster: Handling spatial data

I Function st_read can be used to read vector layers into sf

data structures -
library(sf)

pol_sf = st_read("data/cb_2015_us_state_5m.shp")
## Reading layer `cb_2015_us_state_5m' from data source `/home/michael/Dropbox/
## Simple feature collection with 49 features and 1 field
## geometry type: MULTIPOLYGON
## dimension: XY
## bbox: xmin: -124.7332 ymin: 24.51496 xmax: -66.9499 ymax: 49.38436
## epsg (SRID): 4326
## proj4string: +proj=longlat +datum=WGS84 +no_defs

class(pol_sf)
## [1] "sf" "data.frame"
rgdal, sf, raster: Handling spatial data
I Since sf is new, the majority of the R-Spatial ecosystem only
works with sp
I “Migration” document between sp and sf
I Conversion sp → sf

x = st_as_sf(pol_sp)
class(x)
## [1] "sf" "data.frame"

I Conversion sf → sp

x = as(pol_sf, "Spatial")
class(x)
## [1] "SpatialPolygonsDataFrame"
## attr(,"package")
## [1] "sp"
rgdal, sf, raster: Handling spatial data

I Functions raster, stack and brick can be used to read

single-band rasters into raster data structures -
library(raster)

r = raster("data/dem.tif")
r
## class : RasterLayer
## dimensions : 333, 286, 95238 (nrow, ncol, ncell)
## resolution : 90, 90 (x, y)
## extent : 673414, 699154, 3615239, 3645209 (xmin, xmax, ymin, ymax)
## coord. ref. : +proj=utm +zone=36 +ellps=WGS84 +units=m +no_defs
## data source : /home/michael/Dropbox/Presentations/p_2018_02_EO_Workshop_Base
## names : dem
## values : -12, 537 (min, max)

class(r)
## [1] "RasterLayer"
## attr(,"package")
## [1] "raster"
rgdal, sf, raster: Handling spatial data
I CRS of sp vector layer

proj4string(pol_sp)
## [1] "+proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +t

I CRS of raster layer

proj4string(r)
## [1] "+proj=utm +zone=36 +ellps=WGS84 +units=m +no_defs"

I CRS of sf vector layer

st_crs(pol_sf)
## Coordinate Reference System:
## EPSG: 4326
## proj4string: "+proj=longlat +datum=WGS84 +no_defs"
rgeos, sf: Geoprocessing Vector Layers
I GEOS is used for geometric operations on vector layers with
rgeos / sf -
I Numeric operators - Area, Length, Distance. . .
I Logical operators - Contains, Within, Within distance,
Crosses, Overlaps, Equals, Intersects, Disjoint, Touches. . .
I Geometry generating operators - Centroid, Buffer,
Intersection, Union, Difference, Convex-Hull, Simplification. . .
width: −0.2 width: 0 width: 0.2

width: −0.2 width: 0 width: 0.2

Figure 6: Buffer function

rgeos, sf: Geoprocessing Vector Layers

I Distance using package rgeos

library(rgeos)

gDistance(pol_sp[1, ], pol_sp[2, ], byid = TRUE)

## 0
## 1 20.57536

I Distance using package sf

st_distance(pol_sf[1, ], pol_sf[2, ])
## Units: m
## [,1]
## [1,] 1888200
raster: Geoprocessing Rasters

I Geometric operations on rasters can be done with package

raster -
I Accessing cell values - As vector, As matrix, Extract to points
/ lines / polygons, random / regular sampling, Frequency table,
Histogram. . .
I Raster algebra - Arthmetic (+, -, . . . ), Math (sqrt, log10,
. . . ), logical (!, ==, >, . . . ), summary (mean, max, . . . ), Mask,
Overlay. . .
I Changing resolution and extent - Crop, Mosaic,
(Dis)aggregation, Reprojection, Resampling, Shift, Rotation. . .
I Focal operators - Distance, Direction, Focal Filter, Slope,
Aspect, Flow direction. . .
I Transformations - Vector layers <-> Raster. . .
raster: Geoprocessing Rasters
slope_asp = terrain(
r,
opt = c("slope", "aspect"),
unit = "degrees"
)

Elevation (m) Slope (°) Aspect (°)

500 350
40
300
400
50 30 250
300 200
20
0

20
150

150
10

200
0
35

300 100
0

100 10
50
50

45
250 0 0 0 0
40
0

Figure 7: Topographic aspect and slope calculation

geosphere: Geometric calculations on longitude/latitude
I Package geosphere implements spherical trigonometry
functions for distance and direction-related calculations on
geographic coordinates (lon-lat)
100
50

*
Latitude

0
−50
−100

−150 −100 −50 0 50 100 150

Longitude

Figure 8: Points on Great Circle

geosphere: Geometric calculations on longitude/latitude

Figure 9: Visualizing Facebook Friends with geosphere4

4
http://paulbutler.org/archives/visualizing-facebook-friends/
geosphere: Example

library(geosphere)

LA = c(-118.40, 33.95)
NY = c(-73.78, 40.63)
gci = gcIntermediate(LA, NY)

head(gci)
## lon lat
## [1,] -117.6264 34.23369
## [2,] -116.8477 34.51250
## [3,] -116.0638 34.78633
## [4,] -115.2747 35.05510
## [5,] -114.4804 35.31872
## [6,] -113.6810 35.57712
rmapshaper
I rmapshaper is an R wrapper around the mapshaper tool,
which also has a nice web-interface
I The main advantage of rmapshaper is the topologically-aware
simplification algorithm (also see Vignette)
I Shared boundaries between polygons are always kept intact,
with no gaps or overlaps

library(rmapshaper)
pol_sp_s1 = ms_simplify(pol_sp, keep = 0.05)
pol_sp_s2 = ms_simplify(pol_sp, keep = 0.01)

Figure 10: Original (left), 0.05 points retained (middle) and 0.01 retained
(right)
gstat: Geostatistical Modelling
I Univariate and multivariate geostatistics -
I Variogram modelling
I Ordinary and universal point or block (co)kriging
I Cross-validation

333000

7.0
332000

6.5

6.0
331000

5.5

5.0
330000

178500 179500 180500 181500

Figure 11: Predicted Zinc concentration, using Ordinary Kriging

gstat: Example
I Preparing input data

library(gstat)
library(automap)

data(meuse)
data(meuse.riv)
coordinates(meuse) = ~ x + y
data(meuse.grid)
gridded(meuse.grid) = ~ x + y
grid = raster(meuse.grid)
grid[!is.na(grid)] = 1

I Plot

plot(grid)
plot(meuse, add = TRUE)
gstat: Example

333000
1.0010
332000

1.0005

1.0000
331000

0.9995

0.9990
330000

178500 179500 180500 181500

Figure 12: Area of interest and point measurements

gstat: Example

I Function autofitVariogram in package automap provides an

automatic variogram model fitting procedure

v = autofitVariogram(log(zinc) ~ 1, meuse)

v$var_model
## model psill range
## 1 Nug 0.04848089 0.0000
## 2 Sph 0.58754741 889.9084
gstat: Example
I The fitted variogram model is shown below -

plot(v)

Experimental variogram and fitted variogram model

1314
1139
0.6 1349 1355
0.5 830
Semi−variance

0.4
711
0.3
149
184
0.2
36114
Model: Sph

0.1 17 Nugget: 0.05

Sill: 0.64
Range: 890

500 1000 1500

Distance

Figure 13: Fitted variogram model

gstat: Example

I Function gstat creates a geostatistical model object (class

gstat)
I The object holds all necessary information for making
predictions -
I The model definition (formula, model)
I The calibration data (data)

g = gstat(
formula = log(zinc) ~ 1,
model = v$var_model,
data = meuse
)
gstat: Example

I Interpolation requires the geostatistical model and the new

locations -

predicted = interpolate(grid, g)
## [using ordinary kriging]

I Plotting the result -

plot(predicted)
contour(predicted, add = TRUE)
gstat: Example
7

333000
6.8

6.4
6.2
7
7.2

7.4
332000 7.0
6.8 6.4 5.2
6
5.6 6.5

7
6.0
331000

5
5.5
7
6.6 5.4 5.0
5.8
6
6.2
6.6

6.8
330000

6.2 5.2
6.4
6.4 6 6.2

178500 179500 180500 181500

Figure 14: Predicted Zinc concentration, using Ordinary Kriging

gstat: Example

I We are usually interested in predictions within an area of

interest rather than the entire rectangular extent of the
“template” raster
I The mask function can be used to “erase” the irrelevant area

predicted = mask(predicted, grid)

gstat: Example

6
6.

6.8
333000

4
5.
4
6.
7.0

7
332000
6.8 5.2
6
6.6 5.6 6.5

6.0
331000

5.5
5
6.4
8
6.

5.6 5.4 5.0

6.6
6
6.2 6.2
5.8

6.4
6.6
330000

6
6.2 5.2

6.4 5.8

178500 179500 180500 181500

Figure 15: Predicted Zinc concentration, using Ordinary Kriging

spdep: Spatial dependence modelling

I Modelling with spatial weights -

I Building neighbour lists and spatial weights
I Tests for spatial autocorrelation for areal data (e.g. Moran’s I)
I Spatial regression models (e.g. SAR, CAR)

Figure 16: Neighbours list based on regions with contiguous boundaries

spdep: Example

I Loading the North Carolina dataset, see ?spData::nc.sids

file = system.file("shape/nc.shp", package = "sf")

nc = st_read(file)

I Calculating Sudden Infant Death (SID) rate

nc$rate = nc$SID79 / nc$BIR79

spdep: Example
plot(nc[, "rate"])

rate

0.006
0.004
0.002
0.000
Figure 17: SIDS rate in North Carolina
spdep: Example

I Creating spatial weights is a necessary step in using areal

data -
I Defining which areas are concidered “neighbors”
I Assigning weights to the identified neighbor links
I A common approach is to give a weight of 1/n to all n areas
that share a common boundary with given focal area, and 0 to
all other areas
I Function poly2nb can be used to define the neighbors -

library(spdep)
nc = as(nc, "Spatial")
nb = poly2nb(nc)
spdep: Example

summary(nb)
## Neighbour list object:
## Number of regions: 100
## Number of nonzero links: 490
## Percentage nonzero weights: 4.9
## Average number of links: 4.9
## Link number distribution:
##
## 2 3 4 5 6 7 8 9
## 8 15 17 23 19 14 2 2
## 8 least connected regions:
## 4 21 45 56 77 80 90 99 with 2 links
## 2 most connected regions:
## 39 67 with 9 links
spdep: Example

plot(nc, border = "grey")

plot(nb, coordinates(nc), add = TRUE, col = "black")

Figure 18: Neighbours list based on regions with contiguous boundaries

spdep: Example

I Then function nb2listw can then be used to calculate the

neighbor weights -

nbw = nb2listw(nb)
spdep: Example

summary(nbw)
## Characteristics of weights list object:
## Neighbour list object:
## Number of regions: 100
## Number of nonzero links: 490
## Percentage nonzero weights: 4.9
## Average number of links: 4.9
## Link number distribution:
##
## 2 3 4 5 6 7 8 9
## 8 15 17 23 19 14 2 2
## 8 least connected regions:
## 4 21 45 56 77 80 90 99 with 2 links
## 2 most connected regions:
## 39 67 with 9 links
##
## Weights style: W
## Weights constants summary:
## n nn S0 S1 S2
## W 100 10000 100 44.65023 410.4746
spdep: Example
I Moran’s I is the most common global test for spatial
auto-correlation
I Values of I range from −1 to +1, with the expected value
being −1/(N − 1)
I Low values indicate negative spatial autocorrelation
I High values indicate positive spatial autocorrelation
moran.test(nc$rate, nbw)
##
## Moran I test under randomisation
##
## data: nc$rate
## weights: nbw
##
## Moran I statistic standard deviate = 2.3625, p-value = 0.009075
## alternative hypothesis: greater
## sample estimates:
## Moran I statistic Expectation Variance
## 0.142750422 -0.010101010 0.004185853
Disease mapping

Figure 19: Chapter 10 Disease Mapping, in Bivand et al. 2013

spatstat: Spatial point pattern analysis
I Techniques for statistical analysis of spatial point patterns,
such as -
I Kernel density estimation
I Detection of clustering using Ripley’s K-function
I Spatial logistic regression

0.1 0.14

0.0
4
0.2

0.1

6
0.06

6
0.04 0.1

0.0
0.16
0.04 0.04 0.16
0.06 0.12
2 0.08
0.1 0.06
0.04 0.06
0.1 0.1 0.08
0.04 0.04
0.06 0.04
0.04
0.08
0.06 0.08 0.06
0.04 0.04 0.04 0.06

2
0.1
0.08 0.04
0.04 0.04
0.12 0.06

0.1
0.14

0.06
0.04 0.06
0.06 0.04

0.1
0.04

0.04 0.08 0.04 0.0

0.04 4
0.1 0.04 0.04 0.08 0.06 0.04
0.08
0.06
0.04 0.04
0.04 0.06
0.04 0.06
0.06
0.06 0.04 0.06 0.08 0.06
0.08 0.06
0.08 0.04
0.1

0.06 0.04 0.04

0.04
0.06 0.04
0.04
0.08 0.1
0.06 0.04
0.1
0.06
0.04 0.08
0.06

0.14
0.1 0.04
0.04 0.04 6
0.04
0.0

2
0.1 0.14 0.12
0.1
6
0.06

0.0

6
0.1
0.18 0.08
6

Figure 20: Distance map for the Biological Cells Point Pattern dataset
spatstat: Example

I The spatstat package comes with several sample datasets,

including -
I cells - Biological Cells Point Pattern
I japanesepines - Japanese Pines Point Pattern
I redwood - California Redwoods Point Pattern

library(spatstat)
cells
## Planar point pattern: 42 points
## window: rectangle = [0, 1] x [0, 1] units
japanesepines
## Planar point pattern: 65 points
## window: rectangle = [0, 1] x [0, 1] units (one unit = 5.7 metres)
redwood
## Planar point pattern: 62 points
## window: rectangle = [0, 1] x [-1, 0] units
spatstat: Example

plot(cells)
plot(japanesepines)
plot(redwood)

cells japanesepines redwood

Figure 21: Distance map for the Biological Cells Point Pattern dataset
spatstat: Example
I Ripley’s K-function measures the expected number of
random points within a distance r of a typical random point of
a point pattern X

r = seq(0, sqrt(2)/6, by = 0.005)

Kcells = envelope(
cells,
fun = Kest,
r = r,
nsim = 99
)
## Generating 99 simulations of CSR ...
## 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 1
## 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52,
## 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90,
##
## Done.
spatstat: Example

Kpines = envelope(
japanesepines,
fun = Kest,
r = r,
nsim = 99
)
Kredwood = envelope(
redwood,
fun = Kest,
r = r,
nsim = 99
)
spatstat: Example
plot(Kcells)
plot(Kpines)
plot(Kredwood)

Kcells Kpines Kredwood

0.20

0.20
0.20

K ob s (r ) K ob s (r ) K ob s (r )
^ ^ ^

K t h eo (r ) K t h eo (r ) K t h eo (r )
0.15
K h i (r ) K h i (r ) K h i (r )
^ ^ ^

0.15
0.15

K l o (r ) K l o (r ) K l o (r )
^ ^ ^
0.10
K (r )

K (r )

K (r )

0.10
0.10

0.05

0.05
0.05
0.00

0.00

0.00
0.00 0.05 0.10 0.15 0.20 0.00 0.05 0.10 0.15 0.20 0.00 0.05 0.10 0.15 0.20

r r (one unit = 5.7 metres) r

Figure 22: Distance map for the Biological Cells Point Pattern dataset
osmdata: Access to OpenStreetMap data
I Accessing OpenStreetMap (OSM) data using the Overpass API

library(osmdata)

q = opq(bbox = "Beer-Sheva, Israel")

q = add_osm_feature(q, key = "highway")
dat = osmdata_sf(q)
lines = dat$osm_lines
pol = dat$osm_polygons
pol = st_cast(pol, "MULTILINESTRING")
pol = st_cast(pol, "LINESTRING")
lines = rbind(lines, pol)
lines = lines[, c("osm_id", "highway")]
lines = st_transform(lines, 32636)

plot(lines)
osmdata: Access to OpenStreetMap data

osm_id highway

Figure 23: Beer-Sheva road network

ggplot2, ggmap: Visualization

Figure 24: London cycle hire journeys with ggplot25

5
http://spatial.ly/2012/02/great-maps-ggplot2/
ggplot2, ggmap: Visualization

Figure 25: Crime density by day with ggplot26

6
http://spatial.ly/2012/02/great-maps-ggplot2/
leaflet, mapview: Web mapping

I Packages leaflet and mapview provide methods to produce

interactive maps using the Leaflet JavaScript library
I Package leaflet gives more low-level control
I Package mapview is a wrapper around leaflet, automating
addition of useful features -
I Commonly used basemaps
I Color scales and legends
I Labels
I Popups
mapview: Example

I Function mapview produces an interactive map given a spatial

object
I zcol="..." specifies the attribute used for symbology
I legend=TRUE adds a legend

library(mapview)

states = st_read("data/us-states.geojson")

mapview(states, zcol = "density", legend = TRUE)

mapview: Example

Figure 26: Intractive mapview map

leaflet: Example
I The leaflet package can be used to produce similar
interactive maps, but with finer control
I Defining a color scale function with colorBin

pal = colorBin(
palette = "YlOrRd",
domain = states$density,
bins = c(0, 10, 20, 50, 100, 200, 500, 1000, Inf)
)

I The color scale function accepts a numeric value and returns

a color code -

pal(153)
## [1] "#FD8D3C"
leaflet: Example

I Creating HTML popups

labels = paste0(
"<b>",
states$name,
"</b><br/>",
format(round(states$density, 2), nsmall = 2),
" people / mi²"
) %>%
lapply(htmltools::HTML)
leaflet: Example

I Initiating map object with leaflet

I Setting initial extent with setView
I Adding tile layer with addProviderTiles

leaflet(states) %>%
setView(-96, 37.8, 4) %>%
addProviderTiles("CartoDB.DarkMatterNoLabels") %>%
leaflet: Example
I Adding polygonal layer with addPolygons
addPolygons(
fillColor = ~pal(density),
weight = 2,
opacity = 1,
color = "white",
dashArray = "3",
fillOpacity = 0.7,
highlight = highlightOptions(
weight = 5,
color = "#666",
dashArray = "",
fillOpacity = 0.7,
bringToFront = TRUE
),
label = labels,
labelOptions = labelOptions(
style = list("font-weight" = "normal", padding = "3px 8px"),
textsize = "15px",
direction = "auto"
)
) %>%
leaflet: Example

I Adding legend with addLegend

addLegend(
pal = pal,
values = ~density,
opacity = 0.7,
title = NULL,
position = "bottomright"
)
leaflet: Example

Figure 27: Intractive leaflet map

Other useful packages

I Geometry
I maptools, dggridR
I Routing
I gdistance, stplanr
I Statistics
I dbscan, gwrr, geoR, nlme, FRK
I Visualization
I rasterVis
I Web APIs
I ggmap
Books
I Hierarchical Modeling and Analysis for Spatial Data (1st
ed 2003, 2nd ed. 2014)
I Model-based Geostatistics (2007)
I Applied Spatial Data Analysis with R (1st ed. 2008, 2nd
ed. 2013)
I A Practical Guide for Geostatistical Mapping (2009)
I Spatial Data Analysis in Ecology and Agriculture using R
(2012)
I Displaying Time Series, Spatial, and Space-Time Data
with R (2014)
I Learning R for Geospatial Analysis (2014)
I An Introduction to R for Spatial Analysis and Mapping
(2015)
I Spatial Point Patterns: Methodology and Applications
with R (2015)
I Geocomputation with R (2018)7
7
https://geocompr.robinlovelace.net/
Online courses and tutorials

I Courses
I https://mgimond.github.io/Spatial/index.html
I http://adamwilson.us/RDataScience/index.html
I http://geog.uoregon.edu/bartlein/courses/geog490/index.html
I Tutorials
I http://www.nickeubank.com/gis-in-r/
I http://neondataskills.org/spatial-data-gis/
Thank you for listening!