Spatial Data Analysis with R

Robert J. Hijmans and Aniruddha Ghosh

Jan 02, 2019

 CONTENTS

1 Introduction 1

2 Scale and distance 3

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Scale and resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Zonation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.4 Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4.1 Distance matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4.2 Distance for longitude/latitude coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5 Spatial influence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5.1 Adjacency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5.2 Two nearest neighbours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.5.3 Weights matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.5.4 Spatial influence for polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6 Raster based distance metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.6.1 distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.6.2 cost distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.6.3 resistance distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3 Spatial autocorrelation 17
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1.1 Temporal autocorrelation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1.2 Spatial autocorrelation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2 Example data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3 Adjacent polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.4 Compute Moran’s I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

4 Interpolation 29
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2 Temperature in California . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2.1 9.2 NULL model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2.2 proximity polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.2.3 Nearest neighbour interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2.4 Inverse distance weighted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Calfornia Air Pollution data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.3.1 Data preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.3.2 Fit a variogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.3.3 Ordinary kriging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.3.4 Compare with other methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.3.5 Cross-validate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

i
5 Spatial distribution models 51
5.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.1.1 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.1.2 Predictors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.1.3 Background data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.1.4 Combine presence and background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2 Fit a model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2.1 CART . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2.2 Random Forest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.3 Predict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.3.1 Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.3.2 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.4 Extrapolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.5 Further reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

6 Local regression 73
6.1 California precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
6.2 California House Price Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.3 Summarize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.4 Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.5 Geographicaly Weighted Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.5.1 By county . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.5.2 By grid cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.6 spgwr package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

7 Spatial regression models 89

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.2 Reading & aggregating data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.2.1 Get the data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.3 Basic OLS model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
7.4 Spatial lag model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.5 Spatial error model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.6 Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

8 Point pattern analysis 103

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.2 Basic statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.3 Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
8.4 Distance based measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.5 Spatstat package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

ii
 CHAPTER

ONE

INTRODUCTION

In this section we introduce a number of approaches and techniques that are commonly used in spatial data analysis
and modelling.
Spatial data are mostly like other data. The same general principles apply. But there are few things that are rather
important to consider when using spatial data that are not common with other data types. These are discussed in
Chapters 2 and 3 and include issues of scale and zonation (the modifiable areal unit problem), distance and spatial
autocorrelation.
The other chapters, introduce methods in different areas of spatial data analysis. These include the three classical area
of spatial statistics (point pattern analysis, regression and inference with spatial data, geostatistics (interpolation using
Kriging), as well some other methods (local and global regression and classification with spatial data).
Some of the material presented here is based on examples in the book “Geographic Information Analysis” by David
O’Sullivan and David J. Unwin. This book provides an excellent and very accessible introduction to spatial data
analysis. It has much more depth than what we present here. But the book does not show how to practically implement
the approaches that are discussed — which is the main purpose of this website.
The spatial statistical methods are treated in much more detail in “Applied Spatial Data Analysis with R” by Bivand,
Pebesma and Gómez-Rubio.
This section builds on our Introduction to Spatial Data Manipulation R, that you should read first.

1
Spatial Data Analysis with R

2 Chapter 1. Introduction
 CHAPTER

TWO

SCALE AND DISTANCE

2.1 Introduction
Scale, aggregations, and distance are two key concepts in spatial data analysis that can be tricky to come to grips with.
This chapter first discusses scale and related concepts resolution, aggregation and zonation. The second part of the
chapter discusses distance and adjacency.

2.2 Scale and resolution

The term “scale” is tricky. In its narrow geographic sense, it is the the ratio of a distance on a (paper) map to the actual
distance. So if a distance of 1 cm on map “A” represents 100 m in the real world, the map scale is 1/10,000 (1:10,000
or 10-4). If 1 cm on map “B” represents 10 km in the real world, the scale of that map is 1/1,000,000. The first map
“A” would have relatively large scale (and high resolution) as compared to the second map “B”, that would have a
small scale (and low resolution). It follows that if the size maps “A” and “B” were the same, map “B” would represent
a much larger area (would have a much larger “spatial extent”). For that reason, most people would refer to map “B”
having a “larger scale”. That is technically wrong, but there is not much point in fighting that, and it is simply best
to avoid the term “scale”, and certainly “small scale” and “large scale”, because that technically means the opposite
of what most people think. If you want to use these terms, you should probably use them how they are commonly
understood; unless you are among cartographers, of course.
Now that mapping has become a computer based activity, scale is even more treacherous. You can use the same data
to make maps of different sizes. These would all have a different scale. With digital data, we are more interested
in the “inherent” or “measurement” scale of the data. This is sometimes referred to as “grain” but I use “(spatial)
resolution”. In the case of raster data the notion of resolution is straightforward: it is the size of the cells. For vector
data resolution is not as well defined, and it can vary largely within a data set, but you can think of it as the average
distance between the nodes (coordinate pairs) of the lines or polygons. Point data do not have a resolution, unless
cases that are within a certain distance of each other are merged into a single point (the actual geographic objects
represented by points, actually do cover some area; so the actual average size of those areas could also be a measure
of interest, but it typically is not).
In the digital world it is easy to create a “false resolution”, either by dividing raster cells into 4 or more smaller cells,
or by adding nodes in-between nodes of polygons. Imagine having polygons with soils data for a country. Let’s say
that these polygons cover, on average, an area of 100 * 100 = 10,000 km^2^. You can transfer the soil properties
associated with each polygon, e.g. pH, to a raster with 1 km^2^ spatial resolution; and now might (incorrectly) say
that you have a 1 km^2^ spatial resolution soils map. So we need to distinguish the resolution of the representation
(data) and the resolution of the measurements or estimates. The lowest of the two is the one that matters.
Why does scale/resolution matter?
First of all, different processes have different spatial and temporal scales at which they operate Levin, 1992 — in this
context, scale refers both to “extent” and “resolution”. Processes that operate over a larger extent (e.g., a forest) can
be studied at a larger resolution (trees) whereas processes that operate over a smaller extent (e.g. a tree) may need to
be studied at the level of leaves.

3
Spatial Data Analysis with R

From a practical perspective: it affects our estimates of length and size. For example if you wanted to know the length
of the coastline of Britain, you could use the length of spatial dataset representing that coastline. You could get rather
different numbers depending on the data set used. The higher the resolution of the spatial data, the longer the coastline
would appear to be. This is not just a problem of the representation (the data), also at a theoretical level, one can
argue that the length of the coastline is not defined, as it becomes infinite if your resolution approaches zero. This is
illustrated here
Resolution also affects our understanding of relationships between variables of interest. In terms of data collection this
means that we want data to be at the highest spatial (and temporal) resolution possible (affordable). We can aggregate
our data to lower resolutions, but it is not nearly as easy, or even impossible to correctly disaggregate (“downscale”)
data to a higher resolution.

2.3 Zonation
Geographic data are often aggregated by zones. While we would like to have data at the most granular level that is
possible or meanigful (individuals, households, plots, sites), reality is that we often can only get data that is aggregated.
Rather than having data for individuals, we may have mean values for all inhabitants of a census district. Data on
population, disease, income, or crop yield, is typically available for entire countries, for a number of sub-national units
(e.g. provinces), or a set of raster cells.
The areas used to aggregate data are arbitrary (at least relative to the data of interest). The way the borders of the areas
are drawn (how large, what shape, where) can strongly affect the patterns we see and the outcome of data analysis.
This is sometimes referred to as the “Modifiable Areal Unit Problem” (MAUP). The problem of analyzing aggregated
data is referred to as “Ecological Inference”.
To illustrate the effect of zonation and aggregation, I create a region with 1000 households. For each household we
know where they live and what their annual income is. I then aggregate the data to a set of zones.
The income distribution data

set.seed(0)
xy <- cbind(x=runif(1000, 0, 100), y=runif(1000, 0, 100))
income <- (runif(1000) * abs((xy[,1] - 50) * (xy[,2] - 50))) / 500

Inspect the data, both spatially and non-spatially. The first two plots show that there are many poor people and a few
rich people. The thrird that there is a clear spatial pattern in where the rich and the poor live.

par(mfrow=c(1,3), las=1)
plot(sort(income), col=rev(terrain.colors(1000)), pch=20, cex=.75, ylab='income')
hist(income, main='', col=rev(terrain.colors(10)), xlim=c(0,5), breaks=seq(0,5,0.5))
plot(xy, xlim=c(0,100), ylim=c(0,100), cex=income, col=rev(terrain.
˓→colors(50))[10*(income+1)])

4 Chapter 2. Scale and distance

 Spatial Data Analysis with R

Income inequality is often expressed with the Gini coefficient.

n <- length(income)
G <- (2 * sum(sort(income) * 1:n)/sum(income) - (n + 1)) / n
G
## [1] 0.5814548

For our data set the Gini coefficient is 0.581.

Now assume that the household data was grouped by some kind of census districts. I create different districts, in our
case rectangular raster cells, and compute mean income for each district.

library(raster)
r1 <- raster(ncol=1, nrow=4, xmn=0, xmx=100, ymn=0, ymx=100, crs=NA)
r1 <- rasterize(xy, r1, income, mean)

r2 <- raster(ncol=4, nrow=1, xmn=0, xmx=100, ymn=0, ymx=100, crs=NA)

r2 <- rasterize(xy, r2, income, mean)

r3 <- raster(ncol=2, nrow=2, xmn=0, xmx=100, ymn=0, ymx=100, crs=NA)

r3 <- rasterize(xy, r3, income, mean)

r4 <- raster(ncol=3, nrow=3, xmn=0, xmx=100, ymn=0, ymx=100, crs=NA)

r4 <- rasterize(xy, r4, income, mean)

r5 <- raster(ncol=5, nrow=5, xmn=0, xmx=100, ymn=0, ymx=100, crs=NA)

r5 <- rasterize(xy, r5, income, mean)

r6 <- raster(ncol=10, nrow=10, xmn=0, xmx=100, ymn=0, ymx=100, crs=NA)

r6 <- rasterize(xy, r6, income, mean)

Have a look at the plots of the income distribution and the sub-regional averages.

par(mfrow=c(2,3), las=1)
plot(r1); plot(r2); plot(r3); plot(r4); plot(r5); plot(r6)

2.3. Zonation 5
Spatial Data Analysis with R

It is not surprising to see that the smaller the regions get, the better the real pattern is captured. But in all cases,
the histograms show that we do not capture the full income distribution (compare to the histogram with the data for
individuals).

par(mfrow=c(1,3), las=1)
hist(r4, main='', col=rev(terrain.colors(10)), xlim=c(0,5), breaks=seq(0, 5, 0.5))
hist(r5, main='', col=rev(terrain.colors(10)), xlim=c(0,5), breaks=seq(0, 5, 0.5))
hist(r6, main='', col=rev(terrain.colors(10)), xlim=c(0,5), breaks=seq(0, 5, 0.5))

6 Chapter 2. Scale and distance

 Spatial Data Analysis with R

2.4 Distance
Distance is a numerical description of how far apart things are. It is the most fundamental concept in geography. After
all, Waldo Tobler’s First Law of Geography states that “everything is related to everything else, but near things are
more related than distant things”. But how far away are things? That is not always as easy a question as it seems. Of
course we can compute distance “as the crow flies” but that is often not relevant. Perhaps you need to also consider
national borders, mountains, or other barriers. The distance between A and B may even by asymetric, meaning that it
the distance from A to B is not the same as from B to A (for example, the President of the United States can call me,
but I cannot call him (or her)); or because you go faster when walking downhill than when waling uphill.

2.4.1 Distance matrix

Distances are often described in a “distance matrix”. In a distance matrix we have a number for the distance between
all objects of interest. If the distance is symmetric, we only need to fill half the matrix.
Let’s create a distance matrix from a set of points. We start with a set of points
Set up the data, using x-y coordinates for each point:

A <- c(40, 43)

B <- c(101, 1)
C <- c(111, 54)
D <- c(104, 65)
E <- c(60, 22)
F <- c(20, 2)
pts <- rbind(A, B, C, D, E, F)
pts
## [,1] [,2]
## A 40 43
## B 101 1
## C 111 54
## D 104 65
## E 60 22
## F 20 2

2.4. Distance 7
Spatial Data Analysis with R

Plot the points and labels:

plot(pts, xlim=c(0,120), ylim=c(0,120), pch=20, cex=2, col='red', xlab='X', ylab='Y',

˓→las=1)

text(pts+5, LETTERS[1:6])

You can use the dist function to make a distance matrix with a data set of any dimension.

dis <- dist(pts)

dis
## A B C D E
## B 74.06079
## C 71.84706 53.93515
## D 67.67570 64.07027 13.03840
## E 29.00000 46.06517 60.20797 61.52235
## F 45.61798 81.00617 104.80935 105.00000 44.72136

We can check that for the first point using Pythagoras’ theorem.

sqrt((40-101)^2 + (43-1)^2)
## [1] 74.06079

We can transform a distance matrix into a normal matrix.

D <- as.matrix(dis)
round(D)
## A B C D E F
## A 0 74 72 68 29 46
## B 74 0 54 64 46 81
## C 72 54 0 13 60 105
## D 68 64 13 0 62 105
## E 29 46 60 62 0 45
## F 46 81 105 105 45 0

Distance matrices are used in all kinds of non-geographical applications. For example, they are often used to create
cluster diagrams (dendograms).

8 Chapter 2. Scale and distance

 Spatial Data Analysis with R

Question 4: Show R code to make a cluster dendogram summarizing the distances between these six sites, and plot it.
See ?hclust.

2.4.2 Distance for longitude/latitude coordinates

Now consider that the values in pts were coordinates in degrees (longitude/latitude). Then the cartesian distance
as computed by the dist function would be incorrect. In that case we can use the pointDistance function from the
raster package.
library(raster)
gdis <- pointDistance(pts, lonlat=TRUE)
gdis
## [,1] [,2] [,3] [,4] [,5] [,6]
## [1,] 0 NA NA NA NA NA
## [2,] 7614198 0 NA NA NA NA
## [3,] 5155577 5946748 0 NA NA NA
## [4,] 4581656 7104895 1286094 0 NA NA
## [5,] 2976166 5011592 5536367 5737063 0 NA
## [6,] 4957298 9013726 9894640 9521864 4859627 0

Question 5: What is the unit of the values in gdis?

2.5 Spatial influence

An important step in spatial statistics and modelling is to get a measure of the spatial influence between geographic
objects. This can be expressed as a function of adjacency or (inverse) distance, and is often expressed as a spatial
weights matrix. Influence is of course very complex and cannot really be measured and it can be estimated in many
ways. For example the influence between a set of polyongs (countries) can be expressed as having a shared border
or not (being ajacent); as the “crow-fly” distance between their centroids;or as the lengths of a shared border, and in
other ways.

2.5.1 Adjacency
Adjacency is an important concept in some spatial analysis. In some cases objects are considered ajacent when they
“touch”, e.g. neighboring countries. In can also be based on distance. This is the most common approach when
analyzing point data.
We create an adjacency matrix for the point data analysed above. We define points as “ajacent” if they are within a
distance of 50 from each other. Given that we have the distance matrix D this is easy to do.
a <- D < 50
a
## A B C D E F
## A TRUE FALSE FALSE FALSE TRUE TRUE
## B FALSE TRUE FALSE FALSE TRUE FALSE
## C FALSE FALSE TRUE TRUE FALSE FALSE
## D FALSE FALSE TRUE TRUE FALSE FALSE
## E TRUE TRUE FALSE FALSE TRUE TRUE
## F TRUE FALSE FALSE FALSE TRUE TRUE

In adjacency matrices the diagonal values are often set to NA (we do not consider a point to be adjacent to itself). And
TRUE/FALSE values are commonly stored as 1/0 (this is equivalent, and we can make this change with a simple
trick: multiplication with 1)
diag(a) <- NA
Adj50 <- a * 1
(continues on next page)

2.5. Spatial influence 9

Spatial Data Analysis with R

(continued from previous page)

Adj50
## A B C D E F
## A NA 0 0 0 1 1
## B 0 NA 0 0 1 0
## C 0 0 NA 1 0 0
## D 0 0 1 NA 0 0
## E 1 1 0 0 NA 1
## F 1 0 0 0 1 NA

2.5.2 Two nearest neighbours

What if you wanted to compute the “two nearest neighbours” (or three, or four) adjacency-matrix? Here is how you
can do that. For each row, we first get the column numbers in order of the values in that row (that is, the numbers
indicate how the values are ordered).

cols <- apply(D, 1, order)

# we need to transpose the result
cols <- t(cols)

And then get columns 2 to 3 (why not column 1?)

cols <- cols[, 2:3]

cols
## [,1] [,2]
## A 5 6
## B 5 3
## C 4 2
## D 3 5
## E 1 6
## F 5 1

As we now have the column numbers, we can make the row-column pairs that we want (rowcols).

rowcols <- cbind(rep(1:6, each=2), as.vector(t(cols)))

head(rowcols)
## [,1] [,2]
## [1,] 1 5
## [2,] 1 6
## [3,] 2 5
## [4,] 2 3
## [5,] 3 4
## [6,] 3 2

We use these pairs as indices to change the values in matrix Ak3.

Ak3 <- Adj50 * 0

Ak3[rowcols] <- 1
Ak3
## A B C D E F
## A NA 0 0 0 1 1
## B 0 NA 1 0 1 0
## C 0 1 NA 1 0 0
## D 0 0 1 NA 1 0
## E 1 0 0 0 NA 1
## F 1 0 0 0 1 NA

10 Chapter 2. Scale and distance

 Spatial Data Analysis with R

2.5.3 Weights matrix

Rather than expressing spatial influence as a binary value (adjacent or not), it is often expressed as a continuous value.
The simplest approach is to use inverse distance (the further away, the lower the value).

W <- 1 / D
round(W, 4)
## A B C D E F
## A Inf 0.0135 0.0139 0.0148 0.0345 0.0219
## B 0.0135 Inf 0.0185 0.0156 0.0217 0.0123
## C 0.0139 0.0185 Inf 0.0767 0.0166 0.0095
## D 0.0148 0.0156 0.0767 Inf 0.0163 0.0095
## E 0.0345 0.0217 0.0166 0.0163 Inf 0.0224
## F 0.0219 0.0123 0.0095 0.0095 0.0224 Inf

Such as “spatial weights” matrix is often “row-normalized”, such that the sum of weights for each row in the matrix is
the same. First we get rid if the Inf values by changing them to NA. (Where did the Inf values come from?)

W[!is.finite(W)] <- NA

Then compute the row sums.

rtot <- rowSums(W, na.rm=TRUE)

# this is equivalent to
# rtot <- apply(W, 1, sum, na.rm=TRUE)
rtot
## A B C D E F
## 0.09860117 0.08170418 0.13530597 0.13285878 0.11141516 0.07569154

And divide the rows by their totals and check if they row sums add up to 1.

W <- W / rtot
rowSums(W, na.rm=TRUE)
## A B C D E F
## 1 1 1 1 1 1

The values in the columns do not add up to 1.

colSums(W, na.rm=TRUE)
## A B C D E F
## 0.9784548 0.7493803 1.2204900 1.1794393 1.1559273 0.7163082

2.5.4 Spatial influence for polygons

Above we looked at adjacency for a set of points. Here we look at it for polygons. The difference is that

library(raster)
p <- shapefile(system.file("external/lux.shp", package="raster"))

To find adjacent polygons, we can use the spdep package.

library(spdep)

We use poly2nb to create a “rook’s case” neighbors-list. And from that a neighbors matrix.

wr <- poly2nb(p, row.names=p$ID_2, queen=FALSE)

wr
(continues on next page)

2.5. Spatial influence 11

Spatial Data Analysis with R

(continued from previous page)

## Neighbour list object:
## Number of regions: 12
## Number of nonzero links: 46
## Percentage nonzero weights: 31.94444
## Average number of links: 3.833333
wm <- nb2mat(wr, style='B', zero.policy = TRUE)
dim(wm)
## [1] 12 12

Inspect the content or wr and wm

wr[1:6]
## [[1]]
## [1] 2 4 5
##
## [[2]]
## [1] 1 3 4 5 6 12
##
## [[3]]
## [1] 2 5 9 12
##
## [[4]]
## [1] 1 2
##
## [[5]]
## [1] 1 2 3
##
## [[6]]
## [1] 2 8 12
wm[1:6,1:11]
## [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10] [,11]
## 1 0 1 0 1 1 0 0 0 0 0 0
## 2 1 0 1 1 1 1 0 0 0 0 0
## 3 0 1 0 0 1 0 0 0 1 0 0
## 4 1 1 0 0 0 0 0 0 0 0 0
## 5 1 1 1 0 0 0 0 0 0 0 0
## 6 0 1 0 0 0 0 0 1 0 0 0

Compute the number of neighbors for each area.

i <- rowSums(wm)
i
## 1 2 3 4 5 6 7 12 8 9 10 11
## 3 6 4 2 3 3 3 4 4 3 5 6

Expresses as percentage

round(100 * table(i) / length(i), 1)

## i
## 2 3 4 5 6
## 8.3 41.7 25.0 8.3 16.7

Plot the links between the polygons.

par(mai=c(0,0,0,0))
plot(p, col='gray', border='blue')
(continues on next page)

12 Chapter 2. Scale and distance

 Spatial Data Analysis with R

(continued from previous page)

xy <- coordinates(p)
plot(wr, xy, col='red', lwd=2, add=TRUE)

Now some alternative approaches to compute “spatial influence”.

Distance based:

wd10 <- dnearneigh(xy, 0, 10)

wd25 <- dnearneigh(xy, 0, 25, longlat=TRUE)

Nearest neighbors:

k3 <- knn2nb(knearneigh(xy, k=3, RANN=FALSE))

k6 <- knn2nb(knearneigh(xy, k=6, RANN=FALSE))
## Warning in knearneigh(xy, k = 6, RANN = FALSE): k greater than one-third of
## the number of data points

Lag-two Rook:

wr2 <- wr
for (i in 1:length(wr)) {
lag1 <- wr[[i]]
lag2 <- wr[lag1]
lag2 <- sort(unique(unlist(lag2)))
lag2 <- lag2[!(lag2 %in% c(wr[[i]], i))]
wr2[[i]] <- lag2
}

And now we plot them all using the plotit function.

2.5. Spatial influence 13

Spatial Data Analysis with R

plotit <- function(nb, lab='') {

plot(p, col='gray', border='white')
plot(nb, xy, add=TRUE, pch=20)
text(6.3, 50.1, paste0('(', lab, ')'), cex=1.25)
}

par(mfrow=c(2, 3), mai=c(0,0,0,0))

plotit(wr, 'adjacency')
plotit(wr2, 'lag-2 adj.')
plotit(wd10, '10 km')
plotit(wd25, '25 km')
plotit(k3, 'k=3')
plotit(k6, 'k=6')

14 Chapter 2. Scale and distance

 Spatial Data Analysis with R

2.5. Spatial influence 15

Spatial Data Analysis with R

2.6 Raster based distance metrics

2.6.1 distance
2.6.2 cost distance
2.6.3 resistance distance

16 Chapter 2. Scale and distance

 CHAPTER

THREE

SPATIAL AUTOCORRELATION

3.1 Introduction
Spatial autocorrelation is an important concept in spatial statistics. It is a both a nuisance, as it complicates statistical
tests, and a feature, as it allows for spatial interpolation. Its computation and properties are often misunderstood. This
chapter discusses what it is, and how statistics describing it can be computed.
Autocorrelation (whether spatial or not) is a measure of similarity (correlation) between nearby observations. To
understand spatial autocorrelation, it helps to first consider temporal autocorrelation.

3.1.1 Temporal autocorrelation

If you measure something about the same object over time, for example a persons weight or wealth, it is likely that
two observations that are close to each other in time are also similar in measurement. Say that over a couple of years
your weight went from 50 to 80 kg. It is unlikely that it was 60 kg one day, 50 kg the next and 80 the day after that.
Rather it probably went up gradually, with the occasional tapering off, or even reverse in direction. The same may be
true with your bank account, but that may also have a marked monthly trend. To measure the degree of association
over time, we can compute the correlation of each observation with the next observation.
Let d be a vector of daily observations.

set.seed(0)
d <- sample(100, 10)
d
## [1] 90 27 37 56 88 20 85 96 61 58

Compute auto-correlation.

a <- d[-length(d)]
b <- d[-1]
plot(a, b, xlab='t', ylab='t-1')

17
Spatial Data Analysis with R

cor(a, b)
## [1] -0.2222057

The autocorrelation computed above is very small. Even though this is a random sample, you (almost) never get a
value of zero. We computed the “one-lag” autocorrelation, that is, we compare each value to its immediate neighbour,
and not to other nearby values.
After sorting the numbers in d autocorrelation becomes very strong (unsurprisingly).

d <- sort(d)
d
## [1] 20 27 37 56 58 61 85 88 90 96
a <- d[-length(d)]
b <- d[-1]
plot(a, b, xlab='t', ylab='t-1')

18 Chapter 3. Spatial autocorrelation

 Spatial Data Analysis with R

cor(a, b)
## [1] 0.9530641

The acf function shows autocorrelation computed in a slightly different way for several lags (it is 1 to each point it
self, very high when comparing with the nearest neighbour, and than tapering off).

acf(d)

3.1. Introduction 19
Spatial Data Analysis with R

3.1.2 Spatial autocorrelation

The concept of spatial autocorrelation is an extension of temporal autocorrelation. It is a bit more complicated though.
Time is one-dimensional, and only goes in one direction, ever forward. Spatial objects have (at least) two dimensions
and complex shapes, and it may not be obvious how to determine what is “near”.
Measures of spatial autocorrelation describe the degree two which observations (values) at spatial locations (whether
they are points, areas, or raster cells), are similar to each other. So we need two things: observations and locations.
Spatial autocorrelation in a variable can be exogenous (it is caused by another spatially autocorrelated variable, e.g.
rainfall) or endogenous (it is caused by the process at play, e.g. the spread of a disease).
A commonly used statistic that describes spatial autocorrelation is Moran’s I, and we’ll discuss that here in detail.
Other indices include Geary’s C and, for binary data, the join-count index. The semi-variogram also expresses the
amount of spatial autocorrelation in a data set (see the chapter on interpolation).

3.2 Example data

Read the example data

library(raster)
p <- shapefile(system.file("external/lux.shp", package="raster"))
p <- p[p$NAME_1=="Diekirch", ]
(continues on next page)

20 Chapter 3. Spatial autocorrelation

 Spatial Data Analysis with R

(continued from previous page)

p$value <- c(10, 6, 4, 11, 6)
data.frame(p)
## ID_1 NAME_1 ID_2 NAME_2 AREA value
## 0 1 Diekirch 1 Clervaux 312 10
## 1 1 Diekirch 2 Diekirch 218 6
## 2 1 Diekirch 3 Redange 259 4
## 3 1 Diekirch 4 Vianden 76 11
## 4 1 Diekirch 5 Wiltz 263 6

Let’s say we are interested in spatial autocorrelation in variable “AREA”. If there were spatial autocorrelation, regions
of a similar size would be spatially clustered.
Here is a plot of the polygons. I use the coordinates function to get the centroids of the polygons to place the
labels.

par(mai=c(0,0,0,0))
plot(p, col=2:7)
xy <- coordinates(p)
points(xy, cex=6, pch=20, col='white')
text(p, 'ID_2', cex=1.5)

3.2. Example data 21

Spatial Data Analysis with R

3.3 Adjacent polygons

Now we need to determine which polygons are “near”, and how to quantify that. Here we’ll use adjacency as criterion.
To find adjacent polygons, we can use package ‘spdep’.

library(spdep)
w <- poly2nb(p, row.names=p$Id)
class(w)
## [1] "nb"
summary(w)
## Neighbour list object:
## Number of regions: 5
## Number of nonzero links: 14
## Percentage nonzero weights: 56
## Average number of links: 2.8
## Link number distribution:
##
## 2 3 4
## 2 2 1
## 2 least connected regions:
## 2 3 with 2 links
## 1 most connected region:
## 1 with 4 links

summary(w) tells us something about the neighborhood. The average number of neighbors (adjacent polygons) is
2.8, 3 polygons have 2 neighbors and 1 has 4 neighbors (which one is that?).
For more details we can look at the structure of w.

str(w)
## List of 5
## $ : int [1:3] 2 4 5
## $ : int [1:4] 1 3 4 5
## $ : int [1:2] 2 5
## $ : int [1:2] 1 2
## $ : int [1:3] 1 2 3
## - attr(*, "class")= chr "nb"
## - attr(*, "region.id")= chr [1:5] "0" "1" "2" "3" ...
## - attr(*, "call")= language poly2nb(pl = p, row.names = p$Id)
## - attr(*, "type")= chr "queen"
## - attr(*, "sym")= logi TRUE

Question 1:Explain the meaning of the first 5 lines returned by str(w)

Plot the links between the polygons.

plot(p, col='gray', border='blue', lwd=2)

plot(w, xy, col='red', lwd=2, add=TRUE)

22 Chapter 3. Spatial autocorrelation

 Spatial Data Analysis with R

We can transform w into a spatial weights matrix. A spatial weights matrix reflects the intensity of the geographic
relationship between observations (see previous chapter).
wm <- nb2mat(w, style='B')
wm
## [,1] [,2] [,3] [,4] [,5]
## 0 0 1 0 1 1
## 1 1 0 1 1 1
## 2 0 1 0 0 1
## 3 1 1 0 0 0
## 4 1 1 1 0 0
## attr(,"call")
## nb2mat(neighbours = w, style = "B")

3.4 Compute Moran’s I

Now let’s compute Moran’s index of spatial autocorrelation
$$ I = \frac{n}{\sum_{i=1}^n (y_i - \bar{y})^2} \frac{\sum_{i=1}^n \sum_{j=1}^n w_{ij}(y_i - \bar{y})(y_j -
\bar{y})}{\sum_{i=1}^n \sum_{j=1}^n w_{ij}} $$
Yes, that looks impressive. But it is not much more than an expanded version of the formula to compute the correlation
coefficient. The main thing that was added is the spatial weights matrix.
The number of observations

3.4. Compute Moran’s I 23

Spatial Data Analysis with R

n <- length(p)

Get ‘y’ and ‘ybar’ (the mean value of y)

y <- p$value
ybar <- mean(y)

Now we need $$ (y_i - \bar{y})(y_j - \bar{y}) $$

That is, (yi-ybar)(yj-ybar) for all pairs. I show two methods to get that.
Method 1:

dy <- y - ybar
g <- expand.grid(dy, dy)
yiyj <- g[,1] * g[,2]

Method 2:

yi <- rep(dy, each=n)

yj <- rep(dy)
yiyj <- yi * yj

Make a matrix of the multiplied pairs

pm <- matrix(yiyj, ncol=n)

And multiply this matrix with the weights to set to zero the value for the pairs that are not adjacent.

pmw <- pm * wm
pmw
## [,1] [,2] [,3] [,4] [,5]
## 0 0.00 -3.64 0.00 9.36 -3.64
## 1 -3.64 0.00 4.76 -5.04 1.96
## 2 0.00 4.76 0.00 0.00 4.76
## 3 9.36 -5.04 0.00 0.00 0.00
## 4 -3.64 1.96 4.76 0.00 0.00
## attr(,"call")
## nb2mat(neighbours = w, style = "B")

We now sum the values, to get this bit of Moran’s I:

$$ \sum_{i=1}^n \sum_{j=1}^n w_{ij}(y_i - \bar{y})(y_j - \bar{y}) $$

spmw <- sum(pmw)

spmw
## [1] 17.04

The next step is to divide this value by the sum of weights. That is easy.

smw <- sum(wm)

sw <- spmw / smw

And compute the inverse variance of y

vr <- n / sum(dy^2)

The final step to compute Moran’s I

24 Chapter 3. Spatial autocorrelation

 Spatial Data Analysis with R

MI <- vr * sw
MI
## [1] 0.1728896

This is a simple (but crude) way to estimate the expected value of Moran’s I. That is, the value you would get in the
absence of spatial autocorelation (if the data were spatially random). Of course you never really expect that, but that
is how we do it in statistics. Note that the expected value approaches zero if n becomes large, but that it is not quite
zero for small values of n.

EI <- -1/(n-1)
EI
## [1] -0.25

After doing this ‘by hand’, now let’s use the spdep package to compute Moran’s I and do a significance test. To do this
we need to create a ‘listw’ type spatial weights object (instead of the matrix we used above). To get the same value as
above we use “style=’B’” to use binary (TRUE/FALSE) distance weights.

ww <- nb2listw(w, style='B')

ww
## Characteristics of weights list object:
## Neighbour list object:
## Number of regions: 5
## Number of nonzero links: 14
## Percentage nonzero weights: 56
## Average number of links: 2.8
##
## Weights style: B
## Weights constants summary:
## n nn S0 S1 S2
## B 5 25 14 28 168

Now we can use the moran function. Have a look at ?moran. The function is defined as ‘moran(y, ww, n, Szero(ww))’.
Note the odd arguments n and S0. I think they are odd, because “ww” has that information. Anyway, we supply them
and it works. There probably are cases where it makes sense to use other values.

moran(p$value, ww, n=length(ww$neighbours), S0=Szero(ww))

## $I
## [1] 0.1728896
##
## $K
## [1] 1.432464

#Note that
Szero(ww)
## [1] 14
# is the same as
pmw
## [,1] [,2] [,3] [,4] [,5]
## 0 0.00 -3.64 0.00 9.36 -3.64
## 1 -3.64 0.00 4.76 -5.04 1.96
## 2 0.00 4.76 0.00 0.00 4.76
## 3 9.36 -5.04 0.00 0.00 0.00
## 4 -3.64 1.96 4.76 0.00 0.00
## attr(,"call")
## nb2mat(neighbours = w, style = "B")
sum(pmw==0)
## [1] 11

3.4. Compute Moran’s I 25

Spatial Data Analysis with R

Now we can test for significance. First analytically, using linear regression based logic and assumptions.

moran.test(p$value, ww, randomisation=FALSE)

##
## Moran I test under normality
##
## data: p$value
## weights: ww
##
## Moran I statistic standard deviate = 2.3372, p-value = 0.009714
## alternative hypothesis: greater
## sample estimates:
## Moran I statistic Expectation Variance
## 0.1728896 -0.2500000 0.0327381

Instead of the approach above you should use Monte Carlo simulation. That is the preferred method (in fact, the only
good method). The oay it works that the values are randomly assigned to the polygons, and the Moran’s I is computed.
This is repeated several times to establish a distribution of expected values. The observed value of Moran’s I is then
compared with the simulated distribution to see how likely it is that the observed values could be considered a random
draw.

moran.mc(p$value, ww, nsim=99)

##
## Monte-Carlo simulation of Moran I
##
## data: p$value
## weights: ww
## number of simulations + 1: 100
##
## statistic = 0.17289, observed rank = 96, p-value = 0.04
## alternative hypothesis: greater

Question 2: How do you interpret these results (the significance tests)?

Also try this code, it gives an error: moran.mc(p$value, ww, nsim=999)
Question 3: What is the maximum value we can use for nsim?
We can make a “Moran scatter plot” to visualize spatial autocorrelation. We first get the neighbouring values for each
value.

n <- length(p)
ms <- cbind(id=rep(1:n, each=n), y=rep(y, each=n), value=as.vector(wm * y))

Remove the zeros

ms <- ms[ms[,3] > 0, ]

And compute the average neighbour value

ams <- aggregate(ms[,2:3], list(ms[,1]), FUN=mean)

ams <- ams[,-1]
colnames(ams) <- c('y', 'spatially lagged y')
head(ams)
## y spatially lagged y
## 1 10 7.666667
## 2 6 7.750000
## 3 4 6.000000
(continues on next page)

26 Chapter 3. Spatial autocorrelation

 Spatial Data Analysis with R

(continued from previous page)

## 4 11 8.000000
## 5 6 6.666667

Finally, the plot.

plot(ams)
reg <- lm(ams[,2] ~ ams[,1])
abline(reg, lwd=2)
abline(h=mean(ams[,2]), lt=2)
abline(v=ybar, lt=2)

Note that the slope of the regression line:

coefficients(reg)[2]
## ams[, 1]
## 0.2315341

is almost the same as Moran’s I.

Here is a more direct approach to accomplish the same thing (but hopefully the above makes it clearer how this is
actually computed). Note the row standardisation of the weights matrix:

3.4. Compute Moran’s I 27

Spatial Data Analysis with R

rwm <- mat2listw(wm, style='W')

# Checking if rows add up to 1
mat <- listw2mat(rwm)
apply(mat, 1, sum)[1:15]
## 0 1 2 3 4 <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA>
## 1 1 1 1 1 NA NA NA NA NA NA NA NA NA NA

Now we can plot

moran.plot(y, rwm)

Question 4: Show how to use the ‘geary’ function to compute Geary’s C

Question 5: Write your own Monte Carlo simulation test to compute p-values for Moran’s I, replicating the results
we obtained with the function from spdep. Show a histogram of the simulated values.
Question 6: Write your own Geary C function, by completing the function below

gearyC <- ((n-1)/sum(( "----")\^2)) * sum(wm * (" --- ")\^2) / (2 * sum(wm))

28 Chapter 3. Spatial autocorrelation

 CHAPTER

FOUR

INTERPOLATION

4.1 Introduction
Almost any variable of interest has spatial autocorrelation. That can be a problem in statistical tests, but it is a
very useful feature when we want to predict values at locations where no measurements have been made; as we
can generally safely assume that values at nearby locations will be similar. There are several spatial interpolation
techniques. We show some of them in this chapter.

4.2 Temperature in California

We will be working with temperature data for California. If have not yet done so, first install the rspatial package to
get the data. You may need to install the devtools package first.

if (!require("rspatial")) devtools::install_github('rspatial/rspatial')
## Loading required package: rspatial

Now get the data

library(rspatial)
d <- sp_data('precipitation')
head(d)
## ID NAME LAT LONG ALT JAN FEB MAR APR MAY JUN
## 1 ID741 DEATH VALLEY 36.47 -116.87 -59 7.4 9.5 7.5 3.4 1.7 1.0
## 2 ID743 THERMAL/FAA AIRPORT 33.63 -116.17 -34 9.2 6.9 7.9 1.8 1.6 0.4
## 3 ID744 BRAWLEY 2SW 32.96 -115.55 -31 11.3 8.3 7.6 2.0 0.8 0.1
## 4 ID753 IMPERIAL/FAA AIRPORT 32.83 -115.57 -18 10.6 7.0 6.1 2.5 0.2 0.0
## 5 ID754 NILAND 33.28 -115.51 -18 9.0 8.0 9.0 3.0 0.0 1.0
## 6 ID758 EL CENTRO/NAF 32.82 -115.67 -13 9.8 1.6 3.7 3.0 0.4 0.0
## JUL AUG SEP OCT NOV DEC
## 1 3.7 2.8 4.3 2.2 4.7 3.9
## 2 1.9 3.4 5.3 2.0 6.3 5.5
## 3 1.9 9.2 6.5 5.0 4.8 9.7
## 4 2.4 2.6 8.3 5.4 7.7 7.3
## 5 8.0 9.0 7.0 8.0 7.0 9.0
## 6 3.0 10.8 0.2 0.0 3.3 1.4

Compute annual precipitation

d$prec <- rowSums(d[, c(6:17)])

plot(sort(d$prec), ylab='Annual precipitation (mm)', las=1, xlab='Stations')

29
Spatial Data Analysis with R

Now make a quick map.

library(sp)
dsp <- SpatialPoints(d[,4:3], proj4string=CRS("+proj=longlat +datum=NAD83"))
dsp <- SpatialPointsDataFrame(dsp, d)
CA <- sp_data("counties")

# define groups for mapping

cuts <- c(0,200,300,500,1000,3000)
# set up a palette of interpolated colors
blues <- colorRampPalette(c('yellow', 'orange', 'blue', 'dark blue'))
pols <- list("sp.polygons", CA, fill = "lightgray")
spplot(dsp, 'prec', cuts=cuts, col.regions=blues(5), sp.layout=pols, pch=20, cex=2)

30 Chapter 4. Interpolation
 Spatial Data Analysis with R

Transform longitude/latitude to planar coordinates, using the commonly used coordinate reference system for Califor-
nia (“Teale Albers”) to assure that our interpolation results will align with other data sets we have.

TA <- CRS("+proj=aea +lat_1=34 +lat_2=40.5 +lat_0=0 +lon_0=-120 +x_0=0 +y_0=-4000000

˓→+datum=NAD83 +units=m +ellps=GRS80 +towgs84=0,0,0")

library(rgdal)
dta <- spTransform(dsp, TA)
cata <- spTransform(CA, TA)

4.2.1 9.2 NULL model

We are going to interpolate (estimate for unsampled locations) the precipitation values. The simplest way would be
to take the mean of all observations. We can consider that a “Null-model” that we can compare other approaches to.
We’ll use the Root Mean Square Error (RMSE) as evaluation statistic.

RMSE <- function(observed, predicted) {

sqrt(mean((predicted - observed)^2, na.rm=TRUE))
}

Get the RMSE for the Null-model

null <- RMSE(mean(dsp$prec), dsp$prec)

null
## [1] 435.3217

4.2. Temperature in California 31

Spatial Data Analysis with R

4.2.2 proximity polygons

Proximity polygons can be used to interpolate categorical variables. Another term for this is “nearest neighbour”
interpolation.

library(dismo)
v <- voronoi(dta)
plot(v)

Looks weird. Let’s confine this to California

ca <- aggregate(cata)
vca <- intersect(v, ca)
spplot(vca, 'prec', col.regions=rev(get_col_regions()))

32 Chapter 4. Interpolation
 Spatial Data Analysis with R

Much better. These are polygons. We can ‘rasterize’ the results like this.

r <- raster(cata, res=10000)

vr <- rasterize(vca, r, 'prec')
plot(vr)

4.2. Temperature in California 33

Spatial Data Analysis with R

Now evaluate with 5-fold cross validation.

set.seed(5132015)
kf <- kfold(nrow(dta))

rmse <- rep(NA, 5)

for (k in 1:5) {
test <- dta[kf == k, ]
train <- dta[kf != k, ]
v <- voronoi(train)
p <- extract(v, test)
rmse[k] <- RMSE(test$prec, p$prec)
}
rmse
## [1] 199.0686 187.8069 166.9153 191.0938 238.9696
mean(rmse)
## [1] 196.7708
1 - (mean(rmse) / null)
## [1] 0.5479875

Question 1: Describe what each step in the code chunk above does
Question 2: How does the proximity-polygon approach compare to the NULL model?
Question 3: You would not typically use proximty polygons for rainfall data. For what kind of data would you use
them?

34 Chapter 4. Interpolation
 Spatial Data Analysis with R

4.2.3 Nearest neighbour interpolation

Here we do nearest neighbour interpolation considering multiple (5) neighbours.
We can use the gstat package for this. First we fit a model. ~1 means “intercept only”. In the case of spatial data,
that would be only ‘x’ and ‘y’ coordinates are used. We set the maximum number of points to 5, and the “inverse
distance power” idp to zero, such that all five neighbors are equally weighted

library(gstat)
gs <- gstat(formula=prec~1, locations=dta, nmax=5, set=list(idp = 0))
nn <- interpolate(r, gs)
## [inverse distance weighted interpolation]
nnmsk <- mask(nn, vr)
plot(nnmsk)

Cross validate the result. Note that we can use the predict method to get predictions for the locations of the test
points.

rmsenn <- rep(NA, 5)

for (k in 1:5) {
test <- dta[kf == k, ]
train <- dta[kf != k, ]
gscv <- gstat(formula=prec~1, locations=train, nmax=5, set=list(idp = 0))
p <- predict(gscv, test)$var1.pred
rmsenn[k] <- RMSE(test$prec, p)
(continues on next page)

4.2. Temperature in California 35

Spatial Data Analysis with R

(continued from previous page)

}
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
rmsenn
## [1] 200.6222 190.8336 180.3833 169.9658 237.9067
mean(rmsenn)
## [1] 195.9423
1 - (mean(rmsenn) / null)
## [1] 0.5498908

4.2.4 Inverse distance weighted

A more commonly used method is “inverse distance weighted” interpolation. The only difference with the nearest
neighbour approach is that points that are further away get less weight in predicting a value a location.

library(gstat)
gs <- gstat(formula=prec~1, locations=dta)
idw <- interpolate(r, gs)
## [inverse distance weighted interpolation]
idwr <- mask(idw, vr)
plot(idwr)

36 Chapter 4. Interpolation
 Spatial Data Analysis with R

Question 4: IDW generated rasters tend to have a noticeable artefact. What is that?
Cross validate. We can predict to the locations of the test points

rmse <- rep(NA, 5)

for (k in 1:5) {
test <- dta[kf == k, ]
train <- dta[kf != k, ]
gs <- gstat(formula=prec~1, locations=train)
p <- predict(gs, test)
rmse[k] <- RMSE(test$prec, p$var1.pred)
}
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
rmse
## [1] 215.3319 211.9383 190.0231 211.8308 230.1893
mean(rmse)
## [1] 211.8627
1 - (mean(rmse) / null)
## [1] 0.5133192

Question 5: Inspect the arguments used for and make a map of the IDW model below. What other name could you

4.2. Temperature in California 37

Spatial Data Analysis with R

give to this method (IDW with these parameters)? Why?

gs2 <- gstat(formula=prec~1, locations=dta, nmax=1, set=list(idp=1))

4.3 Calfornia Air Pollution data

We use California Air Pollution data to illustrate geostatistcal (Kriging) interpolation.

4.3.1 Data preparation

We use the airqual dataset to interpolate ozone levels for California (averages for 1980-2009). Use the variable
OZDLYAV (unit is parts per billion). Original data source.
Read the data file. To get easier numbers to read, I multiply OZDLYAV with 1000

x <- sp_data("airqual.csv")
x$OZDLYAV <- x$OZDLYAV * 1000

Create a SpatialPointsDataFrame and transform to Teale Albers. Note the units=km which was needed to fit the
variogram.

library(sp)
coordinates(x) <- ~LONGITUDE + LATITUDE
proj4string(x) <- CRS('+proj=longlat +datum=NAD83')
TA <- CRS("+proj=aea +lat_1=34 +lat_2=40.5 +lat_0=0 +lon_0=-120 +x_0=0 +y_0=-4000000
˓→+datum=NAD83 +units=km +ellps=GRS80")

library(rgdal)
aq <- spTransform(x, TA)

Create an template raster to interpolate to. E.g., given a SpatialPolygonsDataFrame of California, ‘ca’. Coerce that to
a ‘SpatialGrid’ object (a different representation of the same idea)

cageo <- sp_data('counties.rds')

ca <- spTransform(cageo, TA)
r <- raster(ca)
res(r) <- 10 # 10 km if your CRS's units are in km
g <- as(r, 'SpatialGrid')

4.3.2 Fit a variogram

Use gstat to create an emperical variogram ‘v’

library(gstat)
gs <- gstat(formula=OZDLYAV~1, locations=aq)
v <- variogram(gs, width=20)
head(v)
## np dist gamma dir.hor dir.ver id
## 1 1010 11.35040 34.80579 0 0 var1
## 2 1806 30.63737 47.52591 0 0 var1
## 3 2355 50.58656 67.26548 0 0 var1
## 4 2619 70.10411 80.92707 0 0 var1
## 5 2967 90.13917 88.93653 0 0 var1
## 6 3437 110.42302 84.13589 0 0 var1
plot(v)

38 Chapter 4. Interpolation
 Spatial Data Analysis with R

Now, fit a model variogram

fve <- fit.variogram(v, vgm(85, "Exp", 75, 20))

fve
## model psill range
## 1 Nug 21.96600 0.00000
## 2 Exp 85.52957 72.31404
plot(variogramLine(fve, 400), type='l', ylim=c(0,120))
points(v[,2:3], pch=20, col='red')

4.3. Calfornia Air Pollution data 39

Spatial Data Analysis with R

Try a different type (spherical in stead of exponential)

fvs <- fit.variogram(v, vgm(85, "Sph", 75, 20))

fvs
## model psill range
## 1 Nug 25.57019 0.0000
## 2 Sph 72.65881 135.7744
plot(variogramLine(fvs, 400), type='l', ylim=c(0,120) ,col='blue', lwd=2)
points(v[,2:3], pch=20, col='red')

40 Chapter 4. Interpolation
 Spatial Data Analysis with R

Both look pretty good in this case.

Another way to plot the variogram and the model

plot(v, fve)

4.3. Calfornia Air Pollution data 41

Spatial Data Analysis with R

4.3.3 Ordinary kriging

Use variogram fve in a kriging interpolation

k <- gstat(formula=OZDLYAV~1, locations=aq, model=fve)

# predicted values
kp <- predict(k, g)
## [using ordinary kriging]
spplot(kp)

42 Chapter 4. Interpolation
 Spatial Data Analysis with R

# variance
ok <- brick(kp)
ok <- mask(ok, ca)
names(ok) <- c('prediction', 'variance')
plot(ok)

4.3.4 Compare with other methods

Let’s use gstat again to do IDW interpolation. The basic approach first.

library(gstat)
idm <- gstat(formula=OZDLYAV~1, locations=aq)
idp <- interpolate(r, idm)
(continues on next page)

4.3. Calfornia Air Pollution data 43

Spatial Data Analysis with R

(continued from previous page)

## [inverse distance weighted interpolation]
idp <- mask(idp, ca)
plot(idp)

We can find good values for the idw parameters (distance decay and number of neighbours) through optimization. For
simplicity’s sake I do not do that k times here. The optim function may be a bit hard to grasp at first. But the essence
is simple. You provide a function that returns a value that you want to minimize (or maximize) given a number of
unknown parameters. Your provide initial values for these parameters, and optim then searches for the optimal values
(for which the function returns the lowest number).
RMSE <- function(observed, predicted) {
sqrt(mean((predicted - observed)^2, na.rm=TRUE))
}

f1 <- function(x, test, train) {

nmx <- x[1]
idp <- x[2]
if (nmx < 1) return(Inf)
if (idp < .001) return(Inf)
m <- gstat(formula=OZDLYAV~1, locations=train, nmax=nmx, set=list(idp=idp))
p <- predict(m, newdata=test, debug.level=0)$var1.pred
RMSE(test$OZDLYAV, p)
}
(continues on next page)

44 Chapter 4. Interpolation
 Spatial Data Analysis with R

(continued from previous page)

set.seed(20150518)
i <- sample(nrow(aq), 0.2 * nrow(aq))
tst <- aq[i,]
trn <- aq[-i,]
opt <- optim(c(8, .5), f1, test=tst, train=trn)
opt
## $par
## [1] 9.1569933 0.4521968
##
## $value
## [1] 7.28701
##
## $counts
## function gradient
## 41 NA
##
## $convergence
## [1] 0
##
## $message
## NULL

Our optimal IDW model

m <- gstat(formula=OZDLYAV~1, locations=aq, nmax=opt$par[1], set=list(idp=opt$par[2]))

idw <- interpolate(r, m)
## [inverse distance weighted interpolation]
idw <- mask(idw, ca)
plot(idw)

4.3. Calfornia Air Pollution data 45

Spatial Data Analysis with R

A thin plate spline model

library(fields)
m <- Tps(coordinates(aq), aq$OZDLYAV)
tps <- interpolate(r, m)
tps <- mask(tps, idw)
plot(tps)

46 Chapter 4. Interpolation
 Spatial Data Analysis with R

4.3.5 Cross-validate
Cross-validate the three methods (IDW, Ordinary kriging, TPS) and add RMSE weighted ensemble model.
library(dismo)

nfolds <- 5
k <- kfold(aq, nfolds)

ensrmse <- tpsrmse <- krigrmse <- idwrmse <- rep(NA, 5)

for (i in 1:nfolds) {
test <- aq[k!=i,]
train <- aq[k==i,]
m <- gstat(formula=OZDLYAV~1, locations=train, nmax=opt$par[1], set=list(idp=opt
˓→$par[2]))

p1 <- predict(m, newdata=test, debug.level=0)$var1.pred

idwrmse[i] <- RMSE(test$OZDLYAV, p1)

m <- gstat(formula=OZDLYAV~1, locations=train, model=fve)

p2 <- predict(m, newdata=test, debug.level=0)$var1.pred
krigrmse[i] <- RMSE(test$OZDLYAV, p2)

m <- Tps(coordinates(train), train$OZDLYAV)

(continues on next page)

4.3. Calfornia Air Pollution data 47

Spatial Data Analysis with R

(continued from previous page)

p3 <- predict(m, coordinates(test))
tpsrmse[i] <- RMSE(test$OZDLYAV, p3)

w <- c(idwrmse[i], krigrmse[i], tpsrmse[i])

weights <- w / sum(w)
ensemble <- p1 * weights[1] + p2 * weights[2] + p3 * weights[3]
ensrmse[i] <- RMSE(test$OZDLYAV, ensemble)

}
rmi <- mean(idwrmse)
rmk <- mean(krigrmse)
rmt <- mean(tpsrmse)
rms <- c(rmi, rmt, rmk)
rms
## [1] 7.925989 8.816963 7.588549
rme <- mean(ensrmse)
rme
## [1] 7.718896

Question 6: Which method performed best?

We can use the rmse scores to make a weighted ensemble. Let’s look at the maps

weights <- ( rms / sum(rms) )

s <- stack(idw, ok[[1]], tps)
ensemble <- sum(s * weights)

And compare maps.

s <- stack(idw, ok[[1]], tps, ensemble)

names(s) <- c('IDW', 'OK', 'TPS', 'Ensemble')
plot(s)

48 Chapter 4. Interpolation
 Spatial Data Analysis with R

Question 7: Show where the largest difference exist between IDW and OK.
Question 8: Show where the difference between IDW and OK is within the 95% confidence limit of the OK prediction.
Question 9: Can you describe the pattern we are seeing, and speculate about what is causing it?

4.3. Calfornia Air Pollution data 49

Spatial Data Analysis with R

50 Chapter 4. Interpolation
 CHAPTER

FIVE

SPATIAL DISTRIBUTION MODELS

This page shows how you can use the Random Forest algorithm to make spatial predictions. This approach is widely
used, for example to classify remote sensing data into different land cover classes. But here our objective is to predict
the entire range of a species based on a set of locations where it has been observed. As an example, we use the hominid
species Imaginus magnapedum (also known under the vernacular names of “bigfoot” and “sasquatch”). This species
is so hard to find (at least by scientists) that its very existence is commonly denied by the mainstream media! For
more information about this controversy, see the article by Lozier, Aniello and Hickerson: Predicting the distribution
of Sasquatch in western North America: anything goes with ecological niche modelling.
We want to find out
a) What the complete range of the species might be.
b) How good (general) our model is by predicting the range of the Eastern sub-species, with data from the Western
sub-species.
c) Predict where in Mexico the creature is likely to occur.
d) How climate change might affect its distribution.
In this context, this type of analysis is often referred to as ‘species distribution modeling’ or ‘ecological niche model-
ing’. Here is a more in-depth discussion of this technique.

5.1 Data
5.1.1 Observations
if (!require("rspatial")) devtools::install_github('rspatial/rspatial')
library(rspatial)
bf <- sp_data('bigfoot')
dim(bf)
## [1] 3092 3
head(bf)
## lon lat Class
## 1 -142.9000 61.50000 A
## 2 -132.7982 55.18720 A
## 3 -132.8202 55.20350 A
## 4 -141.5667 62.93750 A
## 5 -149.7853 61.05950 A
## 6 -141.3165 62.77335 A

Plot the locations

plot(bf[,1:2], cex=0.5, col='red')

library(maptools)
(continues on next page)

51
Spatial Data Analysis with R

(continued from previous page)

## Checking rgeos availability: TRUE
data(wrld_simpl)
plot(wrld_simpl, add=TRUE)

5.1.2 Predictors
Supervised classification often uses predictor data obtained from satellite remote sensing. But here, as is com-
mon in species distribution modeling, we use climate data. Specifically, we use ‘bioclimatic variables’, see:
http://www.worldclim.org/bioclim

library(raster)
wc <- raster::getData('worldclim', res=10, var='bio')
plot(wc[[c(1, 12)]], nr=2)

52 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

Now extract climate data for the locations of our observations. That is, get data about the climate that the species likes,
apparently.

bfc <- extract(wc, bf[,1:2])

head(bfc)
## bio1 bio2 bio3 bio4 bio5 bio6 bio7 bio8 bio9 bio10 bio11 bio12 bio13
## [1,] -14 102 27 9672 174 -197 371 51 -11 108 -137 973 119
## [2,] 62 55 31 4136 157 -17 174 43 98 118 15 2602 385
## [3,] 62 55 31 4136 157 -17 174 43 98 118 15 2602 385
## [4,] -57 125 23 15138 206 -332 538 127 -129 127 -256 282 67
## [5,] 10 80 25 8308 174 -140 314 66 5 119 -91 532 81
## [6,] -59 128 23 14923 204 -334 538 122 -130 122 -255 322 75
## bio14 bio15 bio16 bio17 bio18 bio19
## [1,] 43 30 332 156 290 210
## [2,] 128 33 953 407 556 721
## [3,] 128 33 953 407 556 721
## [4,] 6 81 163 22 163 27
## [5,] 22 41 215 72 159 117
## [6,] 8 79 183 28 183 32

# Any missing values?

i <- which(is.na(bfc[,1]))
i
## [1] 862 2667
(continues on next page)

5.1. Data 53
Spatial Data Analysis with R

(continued from previous page)

plot(bf[,1:2], cex=0.5, col='blue')
plot(wrld_simpl, add=TRUE)
points(bf[i, ], pch=20, cex=3, col='red')
## Warning in data.matrix(x): NAs introduced by coercion

Here is a plot that illustrates a component of the ecological niche of our species of interest.

plot(bfc[ ,'bio1'] / 10, bfc[, 'bio12'], xlab='Annual mean temperature (C)',

ylab='Annual precipitation (mm)')

54 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

5.1.3 Background data

Normally, one would build a model that would compare the values of the predictor variables as the locations where
something was observed, with those values at the locations where it was not observed. But we do not have data from
a systematic survey that determined presence and absence. We have presence-only data. (And, determining absence is
not that simple. It is here now, it is gone tomorrow).
The common trick to deal with this is to not model presence versus absence, but presence versus a ‘random expecta-
tion’. This random expectation (also referred to as ‘background’, or ‘random-absence’ data) is what you would get
if the species had no preference for any of the predictor variables (or to other variables that are not in the model, but
correlated with the predictor variables).
There is not much point in taking absence data from very far away (tropical Africa or Antarctica). Typically they are
taken from more or less the entire study area for which we have presences data.

library(dismo)
# extent of all points
e <- extent(SpatialPoints(bf[, 1:2]))
e
## class : Extent
## xmin : -156.75
## xmax : -64.4627
## ymin : 25.141
## ymax : 69.5
(continues on next page)

5.1. Data 55
Spatial Data Analysis with R

(continued from previous page)

# 5000 random samples (excluding NA cells) from extent e

set.seed(0)
bg <- sampleRandom(wc, 5000, ext=e)
dim(bg)
## [1] 5000 19
head(bg)
## bio1 bio2 bio3 bio4 bio5 bio6 bio7 bio8 bio9 bio10 bio11 bio12 bio13
## [1,] 138 173 51 5810 310 -27 337 203 134 211 63 610 151
## [2,] -52 102 19 14357 202 -315 517 112 -222 126 -241 485 84
## [3,] -43 108 22 12307 184 -287 471 99 -133 110 -210 848 112
## [4,] 23 119 41 5842 198 -92 290 -36 103 103 -43 1496 249
## [5,] 204 155 46 6625 357 21 336 273 133 282 113 291 44
## [6,] -94 93 17 15087 173 -354 527 90 -279 95 -286 357 55
## bio14 bio15 bio16 bio17 bio18 bio19
## [1,] 9 83 346 40 321 110
## [2,] 18 52 214 57 209 60
## [3,] 39 32 306 135 300 156
## [4,] 26 62 706 118 118 655
## [5,] 7 60 127 27 125 29
## [6,] 11 52 157 38 145 40

5.1.4 Combine presence and background

d <- rbind(cbind(pa=1, bfc), cbind(pa=0, bg))
d <- data.frame(d)
dim(d)
## [1] 8092 20

5.2 Fit a model

Now we have the data to fit a model. But I am going to split the data into East and West. Let’s say I believe these
are actually are different, albeit related, sub-species (The Eastern Sasquatch is darker and less hairy). I am principally
interested in the western sub-species.

de <- d[bf[,1] > -102, ]

dw <- d[bf[,1] <= -102, ]

5.2.1 CART
Let’s first look at a Classification and Regression Trees (CART) model.

library(rpart)
cart <- rpart(pa~., data=dw)
printcp(cart)
##
## Regression tree:
## rpart(formula = pa ~ ., data = dw)
##
## Variables actually used in tree construction:
## [1] bio10 bio12 bio18 bio19 bio4 bio5
##
## Root node error: 762.45/3246 = 0.23489
(continues on next page)

56 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

(continued from previous page)

##
## n= 3246
##
## CP nsplit rel error xerror xstd
## 1 0.419449 0 1.00000 1.00018 0.0089055
## 2 0.147669 1 0.58055 0.59030 0.0148150
## 3 0.047532 2 0.43288 0.44666 0.0169872
## 4 0.028808 3 0.38535 0.40053 0.0171983
## 5 0.013077 4 0.35654 0.37172 0.0162119
## 6 0.012993 6 0.33039 0.36229 0.0164136
## 7 0.010000 7 0.31739 0.34897 0.0163319
plotcp(cart)

plot(cart, uniform=TRUE, main="Regression Tree")

text(cart, cex=.8)

5.2. Fit a model 57

Spatial Data Analysis with R

# text(cart, use.n=TRUE, all=TRUE, cex=.8)

Question 1: Describe the conditions under which you have the highest probability of finding our beloved species?

5.2.2 Random Forest

CART gives us a nice result to look at that can be easily interpreted (as you just illustrated with your answer to Question
1). But the approach suffers from high variance (meaning that the model will be over-fit, it is different each time a
somewhat different datasets are used). Random Forest does not have that problem as much. Above, with CART, we
use regression, let’s do both regression and classification here. First classification.
library(randomForest)
## randomForest 4.6-14
## Type rfNews() to see new features/changes/bug fixes.
# create a factor to indicated that we want classification
fpa <- as.factor(dw[, 'pa'])

Now fit the RandomForest model

crf <- randomForest(dw[, 2:ncol(dw)], fpa)
crf
##
## Call:
## randomForest(x = dw[, 2:ncol(dw)], y = fpa)
(continues on next page)

58 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

(continued from previous page)

## Type of random forest: classification
## Number of trees: 500
## No. of variables tried at each split: 4
##
## OOB estimate of error rate: 8.9%
## Confusion matrix:
## 0 1 class.error
## 0 1870 152 0.0751731
## 1 137 1087 0.1119281
plot(crf)

The variable importance plot shows which variables are most important in fitting the model. This is computing by
randomizing each variable one by one and then computing the decline in model prediction.

varImpPlot(crf)

5.2. Fit a model 59

Spatial Data Analysis with R

Now we use regression, rather than classification. First we tune a parameter.

trf <- tuneRF(dw[, 2:ncol(dw)], dw[, 'pa'])

## Warning in randomForest.default(x, y, mtry = mtryStart, ntree = ntreeTry, :
## The response has five or fewer unique values. Are you sure you want to do
## regression?
## mtry = 6 OOB error = 0.06789353
## Searching left ...
## Warning in randomForest.default(x, y, mtry = mtryCur, ntree = ntreeTry, :
## The response has five or fewer unique values. Are you sure you want to do
## regression?
## mtry = 3 OOB error = 0.06661352
## 0.01885321 0.05
## Searching right ...
## Warning in randomForest.default(x, y, mtry = mtryCur, ntree = ntreeTry, :
## The response has five or fewer unique values. Are you sure you want to do
## regression?
## mtry = 12 OOB error = 0.06947837
## -0.02334312 0.05

60 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

trf
## mtry OOBError
## 3 3 0.06661352
## 6 6 0.06789353
## 12 12 0.06947837
mt <- trf[which.min(trf[,2]), 1]
mt
## [1] 3

Question 2: What did tuneRF help us find? What does the values of mt represent?

rrf <- randomForest(dw[, 2:ncol(d)], dw[, 'pa'], mtry=mt)

## Warning in randomForest.default(dw[, 2:ncol(d)], dw[, "pa"], mtry = mt):
## The response has five or fewer unique values. Are you sure you want to do
## regression?
rrf
##
## Call:
## randomForest(x = dw[, 2:ncol(d)], y = dw[, "pa"], mtry = mt)
## Type of random forest: regression
## Number of trees: 500
## No. of variables tried at each split: 3
##
## Mean of squared residuals: 0.06365263
(continues on next page)

5.2. Fit a model 61

Spatial Data Analysis with R

(continued from previous page)

## % Var explained: 72.9
plot(rrf)

varImpPlot(rrf)

62 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

5.3 Predict
We can use the model to make predictions to any other place for which we have values for the predictor variables. Our
climate data is global so we could find suitable places for bigfoot in Australia. At first I only want to predict to our
study region, which I define as follows.

# Extent of the western points

ew <- extent(SpatialPoints(bf[bf[,1] <= -102, 1:2]))
ew
## class : Extent
## xmin : -156.75
## xmax : -102.3881
## ymin : 30.77722
## ymax : 69.5

5.3.1 Regression
rp <- predict(wc, rrf, ext=ew)
plot(rp)

5.3. Predict 63
Spatial Data Analysis with R

Note that the regression predictions are well-behaved, in the sense that they are between 0 and 1. However, they are
continuous within that range, and if you wanted presence/absence, you would need a threshold. To get the optimal
threshold, you would normally have a hold out data set, but here I used the training data for simplicity.

eva <- evaluate(dw[dw$pa==1, ], dw[dw$pa==0, ], rrf)

eva
## class : ModelEvaluation
## n presences : 1224
## n absences : 2022
## AUC : 0.9996036
## cor : 0.9680315
## max TPR+TNR at : 0.4869443

We can make a ROC plot

plot(eva, 'ROC')

64 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

Find a good threshold to determine presence/absence and plot the prediction.

tr <- threshold(eva)
tr
## kappa spec_sens no_omission prevalence equal_sens_spec
## thresholds 0.4869443 0.4869443 0.4754381 0.3756083 0.5191059
## sensitivity
## thresholds 0.7497767
plot(rp > tr[1, 'spec_sens'])

5.3. Predict 65
Spatial Data Analysis with R

5.3.2 Classification
We can also use the classification Random Forest model to make a prediction.

rc <- predict(wc, crf, ext=ew)

plot(rc)

66 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

You can also get probabilities for the classes

rc2 <- predict(wc, crf, ext=ew, type='prob', index=2)

plot(rc2)

5.3. Predict 67
Spatial Data Analysis with R

5.4 Extrapolation
Now, let’s see if our model is general enough to predict the distribution of the Eastern species.

de <- na.omit(de)
eva2 <- evaluate(de[de$pa==1, ], de[de$pa==0, ], rrf)
eva2
## class : ModelEvaluation
## n presences : 1866
## n absences : 2978
## AUC : 0.4251416
## cor : -0.2611455
## max TPR+TNR at : 8e-04
plot(eva2, 'ROC')

68 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

We can also look at it on a map.

eus <- extent(SpatialPoints(bf[, 1:2]))

eus
## class : Extent
## xmin : -156.75
## xmax : -64.4627
## ymin : 25.141
## ymax : 69.5
rcusa <- predict(wc, rrf, ext=eus)
plot(rcusa)
points(bf[,1:2], cex=.25)

5.4. Extrapolation 69
Spatial Data Analysis with R

Question 4: Why would it be that the model does not extrapolate well?
An important question in the biogeography of the western species is why it does not occur in Mexico. Or if it does,
where would that be?
Let’s see.

mex <- getData('GADM', country='MEX', level=1)

pm <- predict(wc, rrf, ext=mex)
pm <- mask(pm, mex)
plot(pm)

70 Chapter 5. Spatial distribution models

 Spatial Data Analysis with R

Question 5: Where in Mexico are you most likely to encounter western bigfoot?
We can also estimate range shifts due to climate change

fut <- getData('CMIP5', res=10, var='bio', rcp=85, model='AC', year=70)

names(fut)
## [1] "ac85bi701" "ac85bi702" "ac85bi703" "ac85bi704" "ac85bi705"
## [6] "ac85bi706" "ac85bi707" "ac85bi708" "ac85bi709" "ac85bi7010"
## [11] "ac85bi7011" "ac85bi7012" "ac85bi7013" "ac85bi7014" "ac85bi7015"
## [16] "ac85bi7016" "ac85bi7017" "ac85bi7018" "ac85bi7019"
names(wc)
## [1] "bio1" "bio2" "bio3" "bio4" "bio5" "bio6" "bio7" "bio8"
## [9] "bio9" "bio10" "bio11" "bio12" "bio13" "bio14" "bio15" "bio16"
## [17] "bio17" "bio18" "bio19"
names(fut) <- names(wc)
futusa <- predict(fut, rrf, ext=eus, progress='window')
## Loading required namespace: tcltk
plot(futusa)

5.4. Extrapolation 71
Spatial Data Analysis with R

Question 6: Make a map to show where conditions are improving for western bigfoot, and where they are not. Is the
species headed toward extinction?

5.5 Further reading

More on Species distribution modeling with R; and on the use of boosted regression trees in the same context.

72 Chapter 5. Spatial distribution models

 CHAPTER

SIX

LOCAL REGRESSION

Regression models are typically “global”. That is, all date are used simultaneously to fit a single model. In some
cases it can make sense to fit more flexible “local” models. Such models exist in a general regression framework (e.g.
generalized additive models), where “local” refers to the values of the predictor values. In a spatial context local refers
to location. Rather than fitting a single regression model, it is possible to fit several models, one for each location (out
of possibly very many) locations. This technique is sometimes called “geographically weighted regression” (GWR).
GWR is a data exploration technique that allows to understand changes in importance of different variables over space
(which may indicate that the model used is misspecified and can be improved).
There are two examples here. One short example with California precipitation data, and than a more elaborate example
with house price data.

6.1 California precipitation

if (!require("rspatial")) devtools::install_github('rspatial/rspatial')

library(rspatial)
counties <- sp_data('counties')
p <- sp_data('precipitation')
head(p)
## ID NAME LAT LONG ALT JAN FEB MAR APR MAY JUN
## 1 ID741 DEATH VALLEY 36.47 -116.87 -59 7.4 9.5 7.5 3.4 1.7 1.0
## 2 ID743 THERMAL/FAA AIRPORT 33.63 -116.17 -34 9.2 6.9 7.9 1.8 1.6 0.4
## 3 ID744 BRAWLEY 2SW 32.96 -115.55 -31 11.3 8.3 7.6 2.0 0.8 0.1
## 4 ID753 IMPERIAL/FAA AIRPORT 32.83 -115.57 -18 10.6 7.0 6.1 2.5 0.2 0.0
## 5 ID754 NILAND 33.28 -115.51 -18 9.0 8.0 9.0 3.0 0.0 1.0
## 6 ID758 EL CENTRO/NAF 32.82 -115.67 -13 9.8 1.6 3.7 3.0 0.4 0.0
## JUL AUG SEP OCT NOV DEC
## 1 3.7 2.8 4.3 2.2 4.7 3.9
## 2 1.9 3.4 5.3 2.0 6.3 5.5
## 3 1.9 9.2 6.5 5.0 4.8 9.7
## 4 2.4 2.6 8.3 5.4 7.7 7.3
## 5 8.0 9.0 7.0 8.0 7.0 9.0
## 6 3.0 10.8 0.2 0.0 3.3 1.4

plot(counties)
points(p[,c('LONG', 'LAT')], col='red', pch=20)

73
Spatial Data Analysis with R

Compute annual average precipitation

p$pan <- rowSums(p[,6:17])

Global regression model

m <- lm(pan ~ ALT, data=p)

m
##
## Call:
## lm(formula = pan ~ ALT, data = p)
##
## Coefficients:
## (Intercept) ALT
## 523.60 0.17

Create Spatial* objects with a planar crs.

alb <- CRS("+proj=aea +lat_1=34 +lat_2=40.5 +lat_0=0 +lon_0=-120 +x_0=0 +y_0=-4000000

˓→+ellps=GRS80 +datum=NAD83 +units=m +no_defs")

sp <- p
coordinates(sp) = ~ LONG + LAT
crs(sp) <- "+proj=longlat +datum=NAD83"
spt <- spTransform(sp, alb)
ctst <- spTransform(counties, alb)

74 Chapter 6. Local regression

 Spatial Data Analysis with R

Get the optimal bandwidth

library( spgwr )
## NOTE: This package does not constitute approval of GWR
## as a method of spatial analysis; see example(gwr)
bw <- gwr.sel(pan ~ ALT, data=spt)
## Bandwidth: 526221.1 CV score: 64886883
## Bandwidth: 850593.6 CV score: 74209073
## Bandwidth: 325747.9 CV score: 54001118
## Bandwidth: 201848.6 CV score: 44611213
## Bandwidth: 125274.7 CV score: 35746320
## Bandwidth: 77949.39 CV score: 29181737
## Bandwidth: 48700.74 CV score: 22737197
## Bandwidth: 30624.09 CV score: 17457161
## Bandwidth: 19452.1 CV score: 15163436
## Bandwidth: 12547.43 CV score: 19452191
## Bandwidth: 22792.75 CV score: 15512988
## Bandwidth: 17052.67 CV score: 15709960
## Bandwidth: 20218.99 CV score: 15167438
## Bandwidth: 19767.99 CV score: 15156913
## Bandwidth: 19790.05 CV score: 15156906
## Bandwidth: 19781.39 CV score: 15156902
## Bandwidth: 19781.48 CV score: 15156902
## Bandwidth: 19781.47 CV score: 15156902
## Bandwidth: 19781.47 CV score: 15156902
## Bandwidth: 19781.47 CV score: 15156902
## Bandwidth: 19781.47 CV score: 15156902
bw
## [1] 19781.47

Create a regular set of points to estimate parameters for.

r <- raster(ctst, res=10000)
r <- rasterize(ctst, r)
newpts <- rasterToPoints(r)

Run the gwr function

g <- gwr(pan ~ ALT, data=spt, bandwidth=bw, fit.points=newpts[, 1:2])
g
## Call:
## gwr(formula = pan ~ ALT, data = spt, bandwidth = bw, fit.points = newpts[,
## 1:2])
## Kernel function: gwr.Gauss
## Fixed bandwidth: 19781.47
## Fit points: 4087
## Summary of GWR coefficient estimates at fit points:
## Min. 1st Qu. Median 3rd Qu. Max.
## X.Intercept. -702.40121 79.54254 330.48807 735.42718 3468.8702
## ALT -3.91270 0.03058 0.20461 0.41542 4.6133

Link the results back to the raster

slope <- r
intercept <- r
slope[!is.na(slope)] <- g$SDF$ALT
intercept[!is.na(intercept)] <- g$SDF$'(Intercept)'
s <- stack(intercept, slope)
(continues on next page)

6.1. California precipitation 75

Spatial Data Analysis with R

(continued from previous page)

names(s) <- c('intercept', 'slope')
plot(s)

6.2 California House Price Data

We will use house prices data from the 1990 census, taken from “Pace, R.K. and R. Barry, 1997. Sparse Spatial
Autoregressions. Statistics and Probability Letters 33: 291-297.” You can download the data here

houses <- sp_data("houses1990.csv")

dim(houses)
## [1] 20640 9
head(houses)
## houseValue income houseAge rooms bedrooms population households latitude
## 1 452600 8.3252 41 880 129 322 126 37.88
## 2 358500 8.3014 21 7099 1106 2401 1138 37.86
## 3 352100 7.2574 52 1467 190 496 177 37.85
## 4 341300 5.6431 52 1274 235 558 219 37.85
## 5 342200 3.8462 52 1627 280 565 259 37.85
## 6 269700 4.0368 52 919 213 413 193 37.85
## longitude
## 1 -122.23
## 2 -122.22
## 3 -122.24
## 4 -122.25
## 5 -122.25
## 6 -122.25

Each record represents a census “blockgroup”. The longitude and latitude of the centroids of each block group are
available. We can use that to make a map and we can also use these to link the data to other spatial data. For example
to get county-membership of each block group. To do that, let’s first turn this into a SpatialPointsDataFrame to find
out to which county each point belongs.

76 Chapter 6. Local regression

 Spatial Data Analysis with R

library(sp)
coordinates(houses) <- ~longitude+latitude

plot(houses, cex=0.5, pch=1, axes=TRUE)

Now get the county boundaries and assign CRS of the houses data matches that of the counties (because they are both
in longitude/latitude!).

library(raster)
crs(houses) <- crs(counties)

Do a spatial query (points in polygon)

cnty <- over(houses, counties)

head(cnty)
## STATE COUNTY NAME LSAD LSAD_TRANS
## 1 06 001 Alameda 06 County
## 2 06 001 Alameda 06 County
## 3 06 001 Alameda 06 County
## 4 06 001 Alameda 06 County
## 5 06 001 Alameda 06 County
## 6 06 001 Alameda 06 County

6.2. California House Price Data 77

Spatial Data Analysis with R

6.3 Summarize
We can summarize the data by county. First combine the extracted county data with the original data.
hd <- cbind(data.frame(houses), cnty)

Compute the population by county

totpop <- tapply(hd$population, hd$NAME, sum)
totpop
## Alameda Alpine Amador Butte
## 1241779 1113 30039 182120
## Calaveras Colusa Contra Costa Del Norte
## 31998 16275 799017 16045
## El Dorado Fresno Glenn Humboldt
## 128624 662261 24798 116418
## Imperial Inyo Kern Kings
## 108633 18281 528995 91842
## Lake Lassen Los Angeles Madera
## 50631 27214 8721937 88089
## Marin Mariposa Mendocino Merced
## 204241 14302 75061 176457
## Modoc Mono Monterey Napa
## 9678 9956 342314 108030
## Nevada Orange Placer Plumas
## 78510 2340204 170761 19739
## Riverside Sacramento San Benito San Bernardino
## 1162787 1038540 36697 1409740
## San Diego San Francisco San Joaquin San Luis Obispo
## 2425153 683068 477184 203764
## San Mateo Santa Barbara Santa Clara Santa Cruz
## 614816 335177 1486054 216732
## Shasta Sierra Siskiyou Solano
## 147036 3318 43531 337429
## Sonoma Stanislaus Sutter Tehama
## 385296 370821 63689 49625
## Trinity Tulare Tuolumne Ventura
## 13063 309073 48456 649935
## Yolo Yuba
## 138799 58954

Income is harder because we have the median household income by blockgroup. But it can be approximated by first
computing total income by blockgroup, summing that, and dividing that by the total number of households.
# total income
hd$suminc <- hd$income * hd$households
# now use aggregate (similar to tapply)
csum <- aggregate(hd[, c('suminc', 'households')], list(hd$NAME), sum)
# divide total income by number of housefholds
csum$income <- 10000 * csum$suminc / csum$households
# sort
csum <- csum[order(csum$income), ]
head(csum)
## Group.1 suminc households income
## 53 Trinity 11198.985 5156 21720.30
## 58 Yuba 43739.708 19882 21999.65
## 25 Modoc 8260.597 3711 22259.76
## 47 Siskiyou 38769.952 17302 22407.79
(continues on next page)

78 Chapter 6. Local regression

 Spatial Data Analysis with R

(continued from previous page)

## 17 Lake 47612.899 20805 22885.32
## 11 Glenn 20497.683 8821 23237.37
tail(csum)
## Group.1 suminc households income
## 56 Ventura 994094.8 210418 47243.81
## 7 Contra Costa 1441734.6 299123 48198.72
## 30 Orange 3938638.1 800968 49173.48
## 43 Santa Clara 2621895.6 518634 50553.87
## 41 San Mateo 1169145.6 230674 50683.89
## 21 Marin 436808.4 85869 50869.17

6.4 Regression
Before we make a regression model, let’s first add some new variables that we might use, and then see if we can build
a regression model with house price as dependent variable. The authors of the paper used a lot of log tranforms, so
you can also try that.
hd$roomhead <- hd$rooms / hd$population
hd$bedroomhead <- hd$bedrooms / hd$population
hd$hhsize <- hd$population / hd$households

Ordinary least squares regression:

# OLS
m <- glm( houseValue ~ income + houseAge + roomhead + bedroomhead + population,
˓→data=hd)

summary(m)
##
## Call:
## glm(formula = houseValue ~ income + houseAge + roomhead + bedroomhead +
## population, data = hd)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1226134 -48590 -12944 34425 461948
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -6.508e+04 2.533e+03 -25.686 < 2e-16 ***
## income 5.179e+04 3.833e+02 135.092 < 2e-16 ***
## houseAge 1.832e+03 4.575e+01 40.039 < 2e-16 ***
## roomhead -4.720e+04 1.489e+03 -31.688 < 2e-16 ***
## bedroomhead 2.648e+05 6.820e+03 38.823 < 2e-16 ***
## population 3.947e+00 5.081e-01 7.769 8.27e-15 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 6022427437)
##
## Null deviance: 2.7483e+14 on 20639 degrees of freedom
## Residual deviance: 1.2427e+14 on 20634 degrees of freedom
## AIC: 523369
##
## Number of Fisher Scoring iterations: 2
coefficients(m)
(continues on next page)

6.4. Regression 79
Spatial Data Analysis with R

(continued from previous page)

## (Intercept) income houseAge roomhead bedroomhead
## -65075.701407 51786.005862 1831.685266 -47198.908765 264766.186284
## population
## 3.947461

6.5 Geographicaly Weighted Regression

6.5.1 By county
Of course we could make the model more complex, with e.g. squared income, and interactions. But let’s see if we can
do Geographically Weighted regression. One approach could be to use counties.
First I remove records that were outside the county boundaries

hd2 <- hd[!is.na(hd$NAME), ]

Then I write a function to get what I want from the regression (the coefficients in this case)

regfun <- function(x) {

dat <- hd2[hd2$NAME == x, ]
m <- glm(houseValue~income+houseAge+roomhead+bedroomhead+population, data=dat)
coefficients(m)
}

And now run this for all counties using sapply:

countynames <- unique(hd2$NAME)

res <- sapply(countynames, regfun)

Plot of a single coefficient

dotchart(sort(res['income', ]), cex=0.65)

80 Chapter 6. Local regression

 Spatial Data Analysis with R

6.5. Geographicaly Weighted Regression 81

Spatial Data Analysis with R

There clearly is variation in the coefficient ($beta$) for income. How does this look on a map?
First make a data.frame of the results

resdf <- data.frame(NAME=colnames(res), t(res))

head(resdf)
## NAME X.Intercept. income houseAge roomhead
## Alameda Alameda -62373.62 35842.330 591.1001 24147.3182
## Contra Costa Contra Costa -61759.84 43668.442 465.8897 -356.6085
## Alpine Alpine -77605.93 40850.588 5595.4113 NA
## Amador Amador 120480.71 3234.519 -771.5857 37997.0069
## Butte Butte 50935.36 15577.745 -380.5824 9078.9315
## Calaveras Calaveras 91364.72 7126.668 -929.4065 16843.3456
## bedroomhead population
## Alameda 129814.33 8.0570859
## Contra Costa 150662.89 0.8869663
## Alpine NA NA
## Amador -194176.65 0.9971630
## Butte -32272.68 5.7707597
## Calaveras -78749.86 8.8865713

Fix the counties object. There are too many counties because of the presence of islands. I first aggregate (‘dissolve’ in
GIS-speak’) the counties such that a single county becomes a single (multi-)polygon.

dim(counties)
## [1] 68 5
dcounties <- aggregate(counties, vars='NAME')
## Warning in .local(x, ...): Use argument "by" instead of deprecated argument
## "vars"
dim(dcounties)
## [1] 58 1

Now we can merge this SpatialPolygonsDataFrame with data.frame with the regression results.

cnres <- merge(dcounties, resdf, by='NAME')

spplot(cnres, 'income')

82 Chapter 6. Local regression

 Spatial Data Analysis with R

To show all parameters in a ‘conditioning plot’, we need to first scale the values to get similar ranges.

# a copy of the data

cnres2 <- cnres

# scale all variables, except the first one (county name)

# assigning values to a "@data" slot is risky, but (I think) OK here
cnres2@data = data.frame(scale(data.frame(cnres)[, -1]))
spplot(cnres2)

6.5. Geographicaly Weighted Regression 83

Spatial Data Analysis with R

Is this just random noise, or is there spatial autocorrelation?

library(spdep)
nb <- poly2nb(cnres)
plot(cnres)
plot(nb, coordinates(cnres), add=T, col='red')

84 Chapter 6. Local regression

 Spatial Data Analysis with R

lw <- nb2listw(nb)
moran.test(cnres$income, lw)
##
## Moran I test under randomisation
##
## data: cnres$income
## weights: lw
##
## Moran I statistic standard deviate = 2.2473, p-value = 0.01231
## alternative hypothesis: greater
## sample estimates:
## Moran I statistic Expectation Variance
## 0.173419996 -0.017543860 0.007220867
moran.test(cnres$roomhead, lw, na.action=na.omit)
##
## Moran I test under randomisation
##
## data: cnres$roomhead
## weights: lw
## omitted: 2
##
## Moran I statistic standard deviate = 1.3929, p-value = 0.08183
## alternative hypothesis: greater
## sample estimates:
(continues on next page)

6.5. Geographicaly Weighted Regression 85

Spatial Data Analysis with R

(continued from previous page)

## Moran I statistic Expectation Variance
## 0.102596252 -0.017857143 0.007478348

6.5.2 By grid cell

An alternative approach would be to compute a model for grid cells. Let’s use the ‘Teale Albers’ projection (often
used when mapping the entire state of California).

TA <- CRS("+proj=aea +lat_1=34 +lat_2=40.5 +lat_0=0 +lon_0=-120 +x_0=0 +y_0=-4000000

+datum=NAD83 +units=m +no_defs +ellps=GRS80 +towgs84=0,0,0")
countiesTA <- spTransform(counties, TA)

Create a RasteLayer using the extent of the counties, and setting an arbitrary resolution of 50 by 50 km cells

library(raster)
r <- raster(countiesTA)
res(r) <- 50000

Get the xy coordinates for each raster cell:

xy <- xyFromCell(r, 1:ncell(r))

For each cell, we need to select a number of observations, let’s say within 50 km of the center of each cell (thus the
data that are used in different cells overlap). And let’s require at least 50 observations to do a regression.
First transform the houses data to Teale-Albers

housesTA <- spTransform(houses, TA)

crds <- coordinates(housesTA)

Set up a new regression function.

regfun2 <- function(d) {

m <- glm(houseValue~income+houseAge+roomhead+bedroomhead+population, data=d)
coefficients(m)
}

Run the model for al cells if there are at least 50 observations within a radius of 50 km.

res <- list()

for (i in 1:nrow(xy)) {
d <- sqrt((xy[i,1]-crds[,1])^2 + (xy[i,2]-crds[,2])^2)
j <- which(d < 50000)
if (length(j) > 49) {
d <- hd[j,]
res[[i]] <- regfun2(d)
} else {
res[[i]] <- NA
}
}

For each cell get the income coefficient:

inc <- sapply(res, function(x) x['income'])

Use these values in a RasterLayer

86 Chapter 6. Local regression

 Spatial Data Analysis with R

rinc <- setValues(r, inc)

plot(rinc)
plot(countiesTA, add=T)

Moran(rinc)
## [1] 0.3271564

So that was a lot of ‘home-brew-GWR’.

Question 1: Can you comment on weaknesses (and perhaps strengths) of the approaches I have shown?
Question 2: Can you do it the easier and more professional way for these data, using the spgwr package?

6.6 spgwr package

Now use the spgwr package (and the the gwr function) to fit the model. You can do this with all data, as long as you
supply and argument fit.points (to avoid estimating a model for each observation point. You can use a raster
similar to the one I used above (perhaps disaggregate with a factor 2 first).
This is how you can get the points to use:
Create a RasterLayer with the correct extent

6.6. spgwr package 87

Spatial Data Analysis with R

r <- raster(countiesTA)

Set to a desired resolution. I choose 25 km

res(r) <- 25000

I only want cells inside of CA, so I add some more steps.

ca <- rasterize(countiesTA, r)

Extract the coordinates that are not NA.

fitpoints <- rasterToPoints(ca)

I don’t want the third column

fitpoints <- fitpoints[,-3]

Now specificy the model

gwr.model <- ______

gwr returns a list-like object that includes (as first element) a SpatialPointsDataFrame that has the model
coeffients. Plot these using spplot, and after that, transfer them to a RasterBrick object.
To extract the SpatialPointsDataFrame:

sp <- gwr.model$SDF
spplot(sp)

To reconnect these values to the raster structure (etc.)

cells <- cellFromXY(r, fitpoints)

dd <- as.matrix(data.frame(sp))
b <- brick(r, values=FALSE, nl=nrow(dd))
b[cells] <- dd
names(b) <- colnames(dd)
plot(b)

Question 3: Briefly comment on the results and the differences (if any) with the two home-brew examples.

88 Chapter 6. Local regression

 CHAPTER

SEVEN

SPATIAL REGRESSION MODELS

7.1 Introduction
This chapter deals with the problem of inference in (regression) models with spatial data. Inference from regression
models with spatial data can be suspect. In essence this is because nearby things are similar, and it may not be
fair to consider individual cases as independent (they may be pseudo-replicates). Therefore, such models need to be
diagnosed before reporting them. Specifically, it is important to evaluate the for spatial autocorrelation in the residuals
(as these are supposed to be independent, not correlated). If the residuals are spatially autocorrelated, this indicates
that the model is misspecified. In that case you should try to improve the model by adding (and perhaps removing)
important variables. If that is not possible (either because there is no data available, or because you have no clue as
to what variable to look for), you can try formulating a regression model that controls for spatial autocorrelation. We
show some examples of that approach here.

7.2 Reading & aggregating data

We use California house price data from the 2000 Census.

7.2.1 Get the data

if (!require("rspatial")) devtools::install_github('rspatial/rspatial')

library(rspatial)
h <- sp_data('houses2000')

I have selected some variables on on housing and population. You can get more data from the American Fact Finder
http://factfinder2.census.gov (among other web sites).

library(raster)
dim(h)
## [1] 7049 29
names(h)
## [1] "TRACT" "GEOID" "label" "houseValue" "nhousingUn"
## [6] "recHouses" "nMobileHom" "yearBuilt" "nBadPlumbi" "nBadKitche"
## [11] "nRooms" "nBedrooms" "medHHinc" "Population" "Males"
## [16] "Females" "Under5" "MedianAge" "White" "Black"
## [21] "AmericanIn" "Asian" "Hispanic" "PopInHouse" "nHousehold"
## [26] "Families" "householdS" "familySize" "County"

These are the variables we have:

First some data massaging. These are values for Census tracts. I want to analyze these data at the county level. So we
need to aggregate the values.

89
Spatial Data Analysis with R

hh <- aggregate(h, "County")

Now we have the county outlines, but we also need to get the values of interest at the county level. Although it is
possible to do everything in one step in the aggregate function, I prefer to do this step by step. The simplest case is
where we can sum the numbers. For example for the number of houses.

d1 <- data.frame(h)[, c("nhousingUn", "recHouses", "nMobileHom", "nBadPlumbi",

"nBadKitche", "Population", "Males", "Females", "Under5", "White",
"Black", "AmericanIn", "Asian", "Hispanic", "PopInHouse", "nHousehold", "Families")]

d1a <- aggregate(d1, list(County=h$County), sum, na.rm=TRUE)

In other cases we need to use a weighted mean. For example for houseValue

d2 <- data.frame(h)[, c("houseValue", "yearBuilt", "nRooms", "nBedrooms",

"medHHinc", "MedianAge", "householdS", "familySize")]
d2 <- cbind(d2 * h$nHousehold, hh=h$nHousehold)

d2a <- aggregate(d2, list(County=h$County), sum, na.rm=TRUE)

d2a[, 2:ncol(d2a)] <- d2a[, 2:ncol(d2a)] / d2a$hh

Combine these two groups:

d12 <- merge(d1a, d2a, by='County')

And merge the aggregated (from census tract to county level) attribute data with the aggregated polygons

hh <- merge(hh, d12, by='County')

Let’s make some maps, at the orignal Census tract level. We are using a bit more advanced (and slower) plotting
methods here. First the house value, using a legend with 10 intervals.

library(latticeExtra)
## Loading required package: lattice
## Loading required package: RColorBrewer

grps <- 10
brks <- quantile(h$houseValue, 0:(grps-1)/(grps-1), na.rm=TRUE)

p <- spplot(h, "houseValue", at=brks, col.regions=rev(brewer.pal(grps, "RdBu")), col=

˓→"transparent" )

p + layer(sp.polygons(hh))

90 Chapter 7. Spatial regression models

 Spatial Data Analysis with R

This takes very long. spplot (levelplot) is a bit slow when using a large dataset. . .
A map of the median household income.

brks <- quantile(h$medHHinc, 0:(grps-1)/(grps-1), na.rm=TRUE)

p <- spplot(h, "medHHinc", at=brks, col.regions=rev(brewer.pal(grps, "RdBu")), col=

˓→"transparent")

p + layer(sp.polygons(hh))

7.2. Reading & aggregating data 91

Spatial Data Analysis with R

7.3 Basic OLS model

I’ll now make some models with the county-level data. I first compute some new variables (that I might not all use).

hh$fBadP <- pmax(hh$nBadPlumbi, hh$nBadKitche) / hh$nhousingUn

hh$fWhite <- hh$White / hh$Population
hh$age <- 2000 - hh$yearBuilt

f1 <- houseValue ~ age + nBedrooms

m1 <- lm(f1, data=hh)
summary(m1)
##
## Call:
## lm(formula = f1, data = hh)
##
## Residuals:
## Min 1Q Median 3Q Max
## -222541 -67489 -6128 60509 217655
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -628578 233217 -2.695 0.00931 **
## age 12695 2480 5.119 4.05e-06 ***
(continues on next page)

92 Chapter 7. Spatial regression models

 Spatial Data Analysis with R

(continued from previous page)

## nBedrooms 191889 76756 2.500 0.01543 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 94740 on 55 degrees of freedom
## Multiple R-squared: 0.3235, Adjusted R-squared: 0.2989
## F-statistic: 13.15 on 2 and 55 DF, p-value: 2.147e-05

Just for illustration, here is how you can do OLS with matrix algebra. First set up the data. I add a constant variable
‘1’ to X, to get an intercept.

y <- matrix(hh$houseValue)
X <- cbind(1, hh$age, hh$nBedrooms)

Then use matrix algebra

ols <- solve(t(X) %*% X) %*% t(X) %*% y

rownames(ols) <- c('intercept', 'age', 'nBedroom')
ols
## [,1]
## intercept -628577.95
## age 12694.75
## nBedroom 191888.89

So, according to this simple model, “age” is highly significant. The older a house, the more expensive. You pay
1,269,475 dollars more for a house that is 100 years old than a for new house! While the p-value for the number of
bedrooms is not impressive, but every bedroom adds about 200,000 dollars to the value of a house.
Question 1: What would be the price be of a house built in 1999 with three bedrooms?
Let’s see if the errors (model residuals) appear to be randomly distributed in space.

hh$residuals <- residuals(m1)

brks <- quantile(hh$residuals, 0:(grps-1)/(grps-1), na.rm=TRUE)

spplot(hh, "residuals", at=brks, col.regions=rev(brewer.pal(grps, "RdBu")), col="black

˓→")

7.3. Basic OLS model 93

Spatial Data Analysis with R

What do think? Is this random? Let’s see what Mr. Moran would say. First make a neighborhoods list. I add two
links: between San Francisco and Marin County and vice versa (to consider the Golden Gate bridge).

library(spdep)
nb <- poly2nb(hh)
nb[[21]] <- sort(as.integer(c(nb[[21]], 38)))
nb[[38]] <- sort(as.integer(c(21, nb[[38]])))
nb
## Neighbour list object:
## Number of regions: 58
## Number of nonzero links: 278
## Percentage nonzero weights: 8.263971
## Average number of links: 4.793103

par(mai=c(0,0,0,0))
plot(hh)
plot(nb, coordinates(hh), col='red', lwd=2, add=TRUE)

94 Chapter 7. Spatial regression models

 Spatial Data Analysis with R

We can use the neighbour list object to get the average value for the neighbors of each polygon.

resnb <- sapply(nb, function(x) mean(hh$residuals[x]))

cor(hh$residuals, resnb)
## [1] 0.6311218
plot(hh$residuals, resnb, xlab='Residuals', ylab='Mean adjacent residuals')

7.3. Basic OLS model 95

Spatial Data Analysis with R

lw <- nb2listw(nb)

That does not look independent.

moran.mc(hh$residuals, lw, 999)

##
## Monte-Carlo simulation of Moran I
##
## data: hh$residuals
## weights: lw
## number of simulations + 1: 1000
##
## statistic = 0.41428, observed rank = 1000, p-value = 0.001
## alternative hypothesis: greater

Clearly, there is spatial autocorrelation. Our p-values and regression model coefficients cannot be trusted. so let’s try
SAR models.

7.4 Spatial lag model

Here I show a how to do spatial regression with a spatial lag model (lagsarlm)

96 Chapter 7. Spatial regression models

 Spatial Data Analysis with R

m1s = lagsarlm(f1, data=hh, lw, tol.solve=1.0e-30)

summary(m1s)
##
## Call:lagsarlm(formula = f1, data = hh, listw = lw, tol.solve = 1e-30)
##
## Residuals:
## Min 1Q Median 3Q Max
## -108145.2 -49816.3 -1316.3 44604.9 171536.0
##
## Type: lag
## Coefficients: (asymptotic standard errors)
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -418674.1 153693.6 -2.7241 0.006448
## age 5533.6 1698.2 3.2584 0.001120
## nBedrooms 127912.8 50859.7 2.5150 0.011903
##
## Rho: 0.77413, LR test value: 34.761, p-value: 3.7282e-09
## Asymptotic standard error: 0.08125
## z-value: 9.5277, p-value: < 2.22e-16
## Wald statistic: 90.778, p-value: < 2.22e-16
##
## Log likelihood: -727.9964 for lag model
## ML residual variance (sigma squared): 3871700000, (sigma: 62223)
## Number of observations: 58
## Number of parameters estimated: 5
## AIC: 1466, (AIC for lm: 1498.8)
## LM test for residual autocorrelation
## test value: 0.12431, p-value: 0.72441

hh$residuals <- residuals(m1s)

moran.mc(hh$residuals, lw, 999)
##
## Monte-Carlo simulation of Moran I
##
## data: hh$residuals
## weights: lw
## number of simulations + 1: 1000
##
## statistic = -0.016, observed rank = 544, p-value = 0.456
## alternative hypothesis: greater

brks <- quantile(hh$residuals, 0:(grps-1)/(grps-1), na.rm=TRUE)

p <- spplot(hh, "residuals", at=brks, col.regions=rev(brewer.pal(grps, "RdBu")), col=
˓→"transparent")

print( p + layer(sp.polygons(hh)) )

7.4. Spatial lag model 97

Spatial Data Analysis with R

7.5 Spatial error model

And now with a “Spatial error” (or spatial moving average) models (errorsarlm)

m1e = errorsarlm(f1, data=hh, lw, tol.solve=1.0e-30)

summary(m1e)
##
## Call:errorsarlm(formula = f1, data = hh, listw = lw, tol.solve = 1e-30)
##
## Residuals:
## Min 1Q Median 3Q Max
## -100640.7 -47783.1 -2364.5 44180.6 181876.5
##
## Type: error
## Coefficients: (asymptotic standard errors)
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -185443.8 180133.7 -1.0295 0.30325
## age 4313.6 2214.9 1.9475 0.05147
## nBedrooms 117864.4 51564.7 2.2858 0.02227
##
## Lambda: 0.82151, LR test value: 29.781, p-value: 4.8373e-08
## Asymptotic standard error: 0.071111
## z-value: 11.552, p-value: < 2.22e-16
(continues on next page)

98 Chapter 7. Spatial regression models

 Spatial Data Analysis with R

(continued from previous page)

## Wald statistic: 133.46, p-value: < 2.22e-16
##
## Log likelihood: -730.4863 for error model
## ML residual variance (sigma squared): 4.07e+09, (sigma: 63797)
## Number of observations: 58
## Number of parameters estimated: 5
## AIC: 1471, (AIC for lm: 1498.8)

hh$residuals <- residuals(m1e)

moran.mc(hh$residuals, lw, 999)
##
## Monte-Carlo simulation of Moran I
##
## data: hh$residuals
## weights: lw
## number of simulations + 1: 1000
##
## statistic = 0.039033, observed rank = 765, p-value = 0.235
## alternative hypothesis: greater

brks <- quantile(hh$residuals, 0:(grps-1)/(grps-1), na.rm=TRUE)

p <- spplot(hh, "residuals", at=brks, col.regions=rev(brewer.pal(grps, "RdBu")),
col="transparent")
print( p + layer(sp.polygons(hh)) )

7.5. Spatial error model 99

Spatial Data Analysis with R

Are the resiudals spatially autocorrelated for either of these models? Let’s plot them for the spatial error model.

brks <- quantile(hh$residuals, 0:(grps-1)/(grps-1), na.rm=TRUE)

p <- spplot(hh, "residuals", at=brks, col.regions=rev(brewer.pal(grps, "RdBu")),

col="transparent")

print( p + layer(sp.polygons(hh)) )

.png

7.6 Questions
Question 2: The last two maps still seem to show a lot of spatial autocorrelation. But according to the tests there is
none. Now why might that be?
Question 3: One of the most important, or perhaps THE most important aspect of modeling is variable selection. A
misspecified model is never going to be any good, no matter how much you do to, e.g., correct for spatial autocorre-
lation. a) Which variables would you choose from the list?
b) Which new variables could you propose to create from the variables in the list.
c) Which other variables could you add, created from the geometries/location (perhaps other geographic data).
d) add a lot of variables and use stepAIC to select an ‘optimal’ OLS model

100 Chapter 7. Spatial regression models

 Spatial Data Analysis with R

e) check for spatial autocorrelation in the residuals of that model

7.6. Questions 101

Spatial Data Analysis with R

102 Chapter 7. Spatial regression models

 CHAPTER

EIGHT

POINT PATTERN ANALYSIS

8.1 Introduction
We are using a dataset of crimes in a city. Start by reading in the data.

if (!require("rspatial")) devtools::install_github('rspatial/rspatial')
library(rspatial)
city <- sp_data('city')
crime <- sp_data('crime')

Here is a map of both datasets.

plot(city, col='light blue')

points(crime, col='red', cex=.5, pch='+')

A sorted table of the incidence of crime types.

tb <- sort(table(crime$CATEGORY))[-1]
tb
##
## Arson Weapons Robbery
## 9 15 49
(continues on next page)

103
Spatial Data Analysis with R

(continued from previous page)

## Auto Theft Drugs or Narcotics Commercial Burglary
## 86 134 143
## Grand Theft Assaults DUI
## 143 172 212
## Residential Burglary Vehicle Burglary Drunk in Public
## 219 221 232
## Vandalism Petty Theft
## 355 665

Let’s get the coordinates of the crime data, and for this exercise, remove duplicate crime locations. These are the
‘events’ we will use below (later we’ll go back to the full data set).

xy <- coordinates(crime)
dim(xy)
## [1] 2661 2
xy <- unique(xy)
dim(xy)
## [1] 1208 2
head(xy)
## coords.x1 coords.x2
## [1,] 6628868 1963718
## [2,] 6632796 1964362
## [3,] 6636855 1964873
## [4,] 6626493 1964343
## [5,] 6639506 1966094
## [6,] 6640478 1961983

8.2 Basic statistics

Compute the mean center and standard distance for the crime data.

# mean center
mc <- apply(xy, 2, mean)
# standard distance
sd <- sqrt(sum((xy[,1] - mc[1])^2 + (xy[,2] - mc[2])^2) / nrow(xy))

Plot the data to see what we’ve got. I add a summary circle (as in Fig 5.2) by dividing the circle in 360 points and
compute bearing in radians. I do not think this is particularly helpful, but it might be in other cases. And it is always
fun to figure out how to do tis.

plot(city, col='light blue')

points(crime, cex=.5)
points(cbind(mc[1], mc[2]), pch='*', col='red', cex=5)

# make a circle
bearing <- 1:360 * pi/180
cx <- mc[1] + sd * cos(bearing)
cy <- mc[2] + sd * sin(bearing)
circle <- cbind(cx, cy)
lines(circle, col='red', lwd=2)

104 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

8.3 Density
Here is a basic approach to computing point density.

CityArea <- raster::area(city)

dens <- nrow(xy) / CityArea

Question 1a:What is the unit of ‘dens’?

Question 1b:What is the number of crimes per square km?
To compute quadrat counts I first create quadrats (a RasterLayer). I get the extent for the raster from the city polygon,
and then assign an an arbitrary resolution of 1000. (In real life one should always try a range of resolutions, I think).

r <- raster(city)
res(r) <- 1000
r
## class : RasterLayer
## dimensions : 15, 34, 510 (nrow, ncol, ncell)
## resolution : 1000, 1000 (x, y)
## extent : 6620591, 6654591, 1956519, 1971519 (xmin, xmax, ymin, ymax)
## coord. ref. : +proj=lcc +lat_1=38.33333333333334 +lat_2=39.83333333333334 +lat_
˓→0=37.66666666666666 +lon_0=-122 +x_0=2000000 +y_0=500000.0000000001 +datum=NAD83

˓→+units=us-ft +no_defs +ellps=GRS80 +towgs84=0,0,0

To find the cells that are in the city, and for easy display, I create polygons from the RasterLayer.

r <- rasterize(city, r)
plot(r)
quads <- as(r, 'SpatialPolygons')
plot(quads, add=TRUE)
points(crime, col='red', cex=.5)

8.3. Density 105

Spatial Data Analysis with R

The number of events in each quadrat can be counted using the ‘rasterize’ function. That function can be used to
summarize the number of points within each cell, but also to compute statistics based on the ‘marks’ (attributes). For
example we could compute the number of different crime types) by changing the ‘fun’ argument to another function
(see ?rasterize).
nc <- rasterize(coordinates(crime), r, fun='count', background=0)
plot(nc)
plot(city, add=TRUE)

nc has crime counts. As we only have data for the city, the areas outside of the city need to be excluded. We can do
that with the mask function (see ?mask).
ncrimes <- mask(nc, r)
plot(ncrimes)
(continues on next page)

106 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

(continued from previous page)

plot(city, add=TRUE)

Better. Now the frequencies.

f <- freq(ncrimes, useNA='no')

head(f)
## value count
## [1,] 0 48
## [2,] 1 29
## [3,] 2 24
## [4,] 3 22
## [5,] 4 19
## [6,] 5 16
plot(f, pch=20)

8.3. Density 107

Spatial Data Analysis with R

Does this look like a pattern you would have expected? Now compute average number of cases per quadrat.

# number of quadrats
quadrats <- sum(f[,2])
# number of cases
cases <- sum(f[,1] * f[,2])
mu <- cases / quadrats
mu
## [1] 9.261484

And create a table like Table 5.1 on page 130

ff <- data.frame(f)
colnames(ff) <- c('K', 'X')
ff$Kmu <- ff$K - mu
ff$Kmu2 <- ff$Kmu^2
ff$XKmu2 <- ff$Kmu2 * ff$X
head(ff)
## K X Kmu Kmu2 XKmu2
## 1 0 48 -9.261484 85.77509 4117.2042
## 2 1 29 -8.261484 68.25212 1979.3115
## 3 2 24 -7.261484 52.72915 1265.4996
## 4 3 22 -6.261484 39.20618 862.5360
## 5 4 19 -5.261484 27.68321 525.9811
## 6 5 16 -4.261484 18.16025 290.5639

108 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

The observed variance s^2^ is

s2 <- sum(ff$XKmu2) / (sum(ff$X)-1)

s2
## [1] 276.5555

And the VMR is

VMR <- s2 / mu
VMR
## [1] 29.86082

Question 2:What does this VMR score tell us about the point pattern?

8.4 Distance based measures

As we are using a planar coordinate system we can use the dist function to compute the distances between
pairs of points. If we were using longitude/latitude we could compute distance via spherical trigonometry func-
tions. These are available in the sp, raster, and notably the geosphere package (among others). For example, see
raster::pointDistance.

d <- dist(xy)
class(d)
## [1] "dist"

I want to coerce the dist object to a matrix, and ignore distances from each point to itself (the zeros on the diagonal).

dm <- as.matrix(d)
dm[1:5, 1:5]
## 1 2 3 4 5
## 1 0.000 3980.843 8070.429 2455.809 10900.016
## 2 3980.843 0.000 4090.992 6303.450 6929.439
## 3 8070.429 4090.992 0.000 10375.958 2918.349
## 4 2455.809 6303.450 10375.958 0.000 13130.236
## 5 10900.016 6929.439 2918.349 13130.236 0.000
diag(dm) <- NA
dm[1:5, 1:5]
## 1 2 3 4 5
## 1 NA 3980.843 8070.429 2455.809 10900.016
## 2 3980.843 NA 4090.992 6303.450 6929.439
## 3 8070.429 4090.992 NA 10375.958 2918.349
## 4 2455.809 6303.450 10375.958 NA 13130.236
## 5 10900.016 6929.439 2918.349 13130.236 NA

To get, for each point, the minimum distance to another event, we can use the ‘apply’ function. Think of the rows as
each point, and the columns of all other points (vice versa could also work).

dmin <- apply(dm, 1, min, na.rm=TRUE)

head(dmin)
## 1 2 3 4 5 6
## 266.07892 293.58874 47.90260 140.80688 40.06865 510.41231

Now it is trivial to get the mean nearest neighbour distance according to formula 5.5, page 131.

mdmin <- mean(dmin)

8.4. Distance based measures 109

Spatial Data Analysis with R

Do you want to know, for each point, Which point is its nearest neighbour? Use the ‘which.min’ function (but note
that this ignores the possibility of multiple points at the same minimum distance).

wdmin <- apply(dm, 1, which.min)

And what are the most isolated cases? That is the furtest away from their nearest neigbor. I plot the top 25. A bit
complicated.

plot(city)
points(crime, cex=.1)
ord <- rev(order(dmin))

far25 <- ord[1:25]

neighbors <- wdmin[far25]

points(xy[far25, ], col='blue', pch=20)

points(xy[neighbors, ], col='red')

# drawing the lines, easiest via a loop

for (i in far25) {
lines(rbind(xy[i, ], xy[wdmin[i], ]), col='red')
}

Note that some points, but actually not that many, are used as isolated and as a neighbor to an isolated points.
Now on to the G function

max(dmin)
## [1] 1829.738
# get the unique distances (for the x-axis)
distance <- sort(unique(round(dmin)))
# compute how many cases there with distances smaller that each x
Gd <- sapply(distance, function(x) sum(dmin < x))
(continues on next page)

110 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

(continued from previous page)

# normalize to get values between 0 and 1
Gd <- Gd / length(dmin)
plot(distance, Gd)

# using xlim to exclude the extremes

plot(distance, Gd, xlim=c(0,500))

8.4. Distance based measures 111

Spatial Data Analysis with R

Here is a function to show these values in a more standard way.

stepplot <- function(x, y, type='l', add=FALSE, ...) {

x <- as.vector(t(cbind(x, c(x[-1], x[length(x)]))))
y <- as.vector(t(cbind(y, y)))
if (add) {
lines(x,y, ...)
} else {
plot(x,y, type=type, ...)
}
}

And use it for our G function data.

stepplot(distance, Gd, type='l', lwd=2, xlim=c(0,500))

112 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

The steps are so small in our data, that you hardly see the difference.
I use the centers of previously defined raster cells to compute the F function.

# get the centers of the 'quadrats' (raster cells)

p <- rasterToPoints(r)
# compute distance from all crime sites to these cell centers
d2 <- pointDistance(p[,1:2], xy, longlat=FALSE)

# the remainder is similar to the G function

Fdistance <- sort(unique(round(d2)))
mind <- apply(d2, 1, min)
Fd <- sapply(Fdistance, function(x) sum(mind < x))
Fd <- Fd / length(mind)
plot(Fdistance, Fd, type='l', lwd=2, xlim=c(0,3000))

8.4. Distance based measures 113

Spatial Data Analysis with R

Compute the expected distributon (5.12 on page 145)

ef <- function(d, lambda) {

E <- 1 - exp(-1 * lambda * pi * d^2)
}
expected <- ef(0:2000, dens)

Now, let’s combine F and G on one plot.

plot(distance, Gd, type='l', lwd=2, col='red', las=1,

ylab='F(d) or G(d)', xlab='Distance', yaxs="i", xaxs="i")
lines(Fdistance, Fd, lwd=2, col='blue')
lines(0:2000, expected, lwd=2)

legend(1200, .3,
c(expression(italic("G")["d"]), expression(italic("F")["d"]), 'expected'),
lty=1, col=c('red', 'blue', 'black'), lwd=2, bty="n")

114 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

Question 3: What does this plot suggest about the point pattern?
Finally, let’s compute K. Note that I use the original distance matrix ‘d’ here.

distance <- seq(1, 30000, 100)

Kd <- sapply(distance, function(x) sum(d < x)) # takes a while
Kd <- Kd / (length(Kd) * dens)
plot(distance, Kd, type='l', lwd=2)

8.4. Distance based measures 115

Spatial Data Analysis with R

Question 4: Create a single random pattern of events for the city, with the same number of events as the crime data
(object xy). Use function ‘spsample’
Question 5: Compute the G function, and plot it on a single plot, together with the G function for the observed crime
data, and the theoretical expectation (formula 5.12).
Question 6: (Difficult!) Do a Monte Carlo simulation (page 149) to see if the ‘mean nearest distance’ of the observed
crime data is significantly different from a random pattern. Use a ‘for loop’. First write ‘pseudo-code’. That is, say in
natural language what should happen. Then try to write R code that implements this.

8.5 Spatstat package

Above we did some ‘home-brew’ point pattern analysis, we will now use the spatstat package. In research you
would normally use spatstat rather than your own functions, at least for standard analysis. I showed how you make
some of these functions in the previous sections, because understanding how to go about that may allow you to take
things in directions that others have not gone. The good thing about spatstat is that it very well documented (see
http://spatstat.github.io/). The bad thing is that it uses an entirly different sets of classes (ways to represent spatial
data) that we we will use in all other labs (classes from sp and raster); but it is not hard to get used to that.

library(spatstat)

We start with making make a Kernel Density raster. I first create a ‘ppp’ (point pattern) object, as defined in the spatstat
package.

116 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

A ppp object has the coordinates of the points and the analysis ‘window’ (study region). To assign the points locations
we need to extract the coordinates from our SpatialPoints object. To set the window, we first need to to coerce our
SpatialPolygons into an ‘owin’ object. We need a function from the maptools package for this coercion.
Coerce from SpatialPolygons to an object of class “owin” (observation window)

library(maptools)
cityOwin <- as.owin(city)
class(cityOwin)
## [1] "owin"
cityOwin
## window: polygonal boundary
## enclosing rectangle: [6620591, 6654380] x [1956729.8, 1971518.9] units

Extract coordinates from SpatialPointsDataFrame:

pts <- coordinates(crime)

head(pts)
## coords.x1 coords.x2
## [1,] 6628868 1963718
## [2,] 6632796 1964362
## [3,] 6636855 1964873
## [4,] 6626493 1964343
## [5,] 6639506 1966094
## [6,] 6640478 1961983

Now we can create a ‘ppp’ (point pattern) object

p <- ppp(pts[,1], pts[,2], window=cityOwin)

## Warning: 20 points were rejected as lying outside the specified window
## Warning: data contain duplicated points
class(p)
## [1] "ppp"
p
## Planar point pattern: 2641 points
## window: polygonal boundary
## enclosing rectangle: [6620591, 6654380] x [1956729.8, 1971518.9] units
## *** 20 illegal points stored in attr(,"rejects") ***
plot(p)
## Warning in plot.ppp(p): 20 illegal points also plotted

8.5. Spatstat package 117

Spatial Data Analysis with R

Note the warning message about ‘illegal’ points. Do you see them and do you understand why they are illegal?
Having all the data well organized, it is now easy to compute Kernel Density

ds <- density(p)
class(ds)
## [1] "im"
plot(ds, main='crime density')

118 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

Density is the number of points per unit area. Let’s ceck if the numbers makes sense, by adding them up and mulit-
plying with the area of the raster cells. I use raster package functions for that.

nrow(pts)
## [1] 2661
r <- raster(ds)
s <- sum(values(r), na.rm=TRUE)
s * prod(res(r))
## [1] 2640.556

Looks about right. We can also get the information directly from the “im” (image) object

str(ds)
## List of 10
## $ v : num [1:128, 1:128] NA NA NA NA NA NA NA NA NA NA ...
## $ dim : int [1:2] 128 128
## $ xrange: num [1:2] 6620591 6654380
## $ yrange: num [1:2] 1956730 1971519
## $ xstep : num 264
## $ ystep : num 116
## $ xcol : num [1:128] 6620723 6620987 6621251 6621515 6621779 ...
## $ yrow : num [1:128] 1956788 1956903 1957019 1957134 1957250 ...
## $ type : chr "real"
## $ units :List of 3
(continues on next page)

8.5. Spatstat package 119

Spatial Data Analysis with R

(continued from previous page)

## ..$ singular : chr "unit"
## ..$ plural : chr "units"
## ..$ multiplier: num 1
## ..- attr(*, "class")= chr "unitname"
## - attr(*, "class")= chr "im"
## - attr(*, "sigma")= num 1849
## - attr(*, "kernel")= chr "gaussian"
sum(ds$v, na.rm=TRUE) * ds$xstep * ds$ystep
## [1] 2640.556
p$n
## [1] 2641

Here’s another, lenghty, example of generalization. We can interpolate population density from (2000) census data;
assigning the values to the centroid of a polygon (as explained in the book, but not a great technique). We use a
shapefile with census data.

census <- sp_data("census2000.rds")

To compute population density for each census block, we first need to get the area of each polygon. I transform density
from persons per feet^2^ to persons per mile^2^, and then compute population density from POP2000 and the area

census$area <- area(census)

census$area <- census$area/27878400
census$dens <- census$POP2000 / census$area

Now to get the centroids of the census blocks we can use the ‘coordinates’ function again. Note that it actually does
something quite different (with a SpatialPolygons* object) then in the case above (with a SpatialPoints* object).

p <- coordinates(census)
head(p)
## [,1] [,2]
## 0 6666671 1991720
## 1 6655379 1986903
## 2 6604777 1982474
## 3 6612242 1981881
## 4 6613488 1986776
## 5 6616743 1986446

To create the ‘window’ we dissolve all polygons into a single polygon.

win <- aggregate(census)

Let’s look at what we have:

plot(census)
points(p, col='red', pch=20, cex=.25)
plot(win, add=TRUE, border='blue', lwd=3)

120 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

Now we can use ‘Smooth.ppp’ to interpolate. Population density at the points is referred to as the ‘marks’

owin <- as.owin(win)

pp <- ppp(p[,1], p[,2], window=owin, marks=census$dens)
## Warning: 1 point was rejected as lying outside the specified window
pp
## Marked planar point pattern: 645 points
## marks are numeric, of storage type 'double'
## window: polygonal boundary
## enclosing rectangle: [6576938, 6680926] x [1926586.1, 2007558.2] units
## *** 1 illegal point stored in attr(,"rejects") ***

Note the warning message: “1 point was rejected as lying outside the specified window”. That is odd, there is a
polygon that has a centroid that is outside of the polygon. This can happen with, e.g., kidney shaped polygons.
Let’s find and remove this point that is outside the study area.

sp <- SpatialPoints(p, proj4string=CRS(proj4string(win)))

library(rgeos)
i <- gIntersects(sp, win, byid=TRUE)
which(!i)
## [1] 588

Let’s see where it is:

8.5. Spatstat package 121

Spatial Data Analysis with R

plot(census)
points(sp)
points(sp[!i,], col='red', cex=3, pch=20)

You can zoom in using the code below. After running the next line, click on your map twice to zoom to the red dot,
otherwise you cannot continue:
zoom(census)
And add the red points again
points(sp[!i,], col='red')
To only use points that intersect with the window polygon, that is, where ‘i == TRUE’:

pp <- ppp(p[i,1], p[i,2], window=owin, marks=census$dens[i])

plot(pp)
plot(city, add=TRUE)

122 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

And to get a smooth interpolation of population density.

s <- Smooth.ppp(pp)
## Warning: Cross-validation criterion was minimised at right-hand end of
## interval [89.7, 3350]; use arguments hmin, hmax to specify a wider interval
plot(s)
plot(city, add=TRUE)

8.5. Spatstat package 123

Spatial Data Analysis with R

Population density could establish the “population at risk” (to commit a crime) for certain crimes, but not for others.
Maps with the city limits and the incidence of ‘auto-theft’, ‘drunk in public’, ‘DUI’, and ‘Arson’.

par(mfrow=c(2,2), mai=c(0.25, 0.25, 0.25, 0.25))

for (offense in c("Auto Theft", "Drunk in Public", "DUI", "Arson")) {
plot(city, col='grey')
acrime <- crime[crime$CATEGORY == offense, ]
points(acrime, col = "red")
title(offense)
}

124 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

Create a marked point pattern object (ppp) for all crimes. It is important to coerce the marks to a factor variable.

crime$fcat <- as.factor(crime$CATEGORY)

w <- as.owin(city)
xy <- coordinates(crime)
mpp <- ppp(xy[,1], xy[,2], window = w, marks=crime$fcat)
## Warning: 20 points were rejected as lying outside the specified window
## Warning: data contain duplicated points

We can split the mpp object by category (crime)

spp <- split(mpp)

plot(spp[1:4], main='')

8.5. Spatstat package 125

Spatial Data Analysis with R

The crime density by category:

plot(density(spp[1:4]), main='')

And produce K-plots (with an envelope) for ‘drunk in public’ and ‘Arson’. Can you explain what they mean?

spatstat.options(checksegments = FALSE)
ktheft <- Kest(spp$"Auto Theft")
ketheft <- envelope(spp$"Auto Theft", Kest)
(continues on next page)

126 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

(continued from previous page)

## Generating 99 simulations of CSR ...
## 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
˓→24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38,

## 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59,
˓→ 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76,

## 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97,
˓→ 98, 99.
##
## Done.
ktheft <- Kest(spp$"Arson")
ketheft <- envelope(spp$"Arson", Kest)
## Generating 99 simulations of CSR ...
## 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
˓→24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38,

## 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59,
˓→ 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76,

## 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97,
˓→ 98, 99.
##
## Done.

par(mfrow=c(1,2))
plot(ktheft)
plot(ketheft)

Let’s try to answer the question you have been wanting to answer all along. Is population density a good predictor of
being (booked for) “drunk in public” and for “Arson”? One approach is to do a Kolmogorov-Smirnov (‘kstest’) on
‘Drunk in Public’ and ‘Arson’, using population density as a covariate:

KS.arson <- cdf.test(spp$Arson, ds)

KS.arson
##
## Spatial Kolmogorov-Smirnov test of CSR in two dimensions
##
(continues on next page)

8.5. Spatstat package 127

Spatial Data Analysis with R

(continued from previous page)

## data: covariate 'ds' evaluated at points of 'spp$Arson'
## and transformed to uniform distribution under CSR
## D = 0.51463, p-value = 0.009723
## alternative hypothesis: two-sided
KS.drunk <- cdf.test(spp$'Drunk in Public', ds)
KS.drunk
##
## Spatial Kolmogorov-Smirnov test of CSR in two dimensions
##
## data: covariate 'ds' evaluated at points of 'spp$"Drunk in Public"'
## and transformed to uniform distribution under CSR
## D = 0.54119, p-value < 2.2e-16
## alternative hypothesis: two-sided

Question 7: Why is the result surprising, or not surprising?

We can also compare the patterns for “drunk in public” and for “Arson” with the KCross function.

kc <- Kcross(mpp, i = "Drunk in Public", j = "Arson")

ekc <- envelope(mpp, Kcross, nsim = 50, i = "Drunk in Public", j = "Arson")
## Generating 50 simulations of CSR ...
## 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
˓→24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38,

## 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50.
##
## Done.
plot(ekc)

128 Chapter 8. Point pattern analysis

 Spatial Data Analysis with R

Much more about point pattern analysis with spatstat is available here

8.5. Spatstat package 129