plotmap: Plot Maps

In R2BayesX: Estimate Structured Additive Regression Models with 'BayesX'

Plot Maps

Description

The function takes a list polygons and draws the corresponding map. Different colors for each polygon can be used. Typically used for objects of class "mrf.bayesx" and "random.bayesx" returned from function bayesx and read.bayesx.output.

Usage

plotmap(map, x = NULL, id = NULL, c.select = NULL, legend = TRUE,
  missing = TRUE, swap = FALSE, range = NULL, names = FALSE,
  values = FALSE, col = NULL, ncol = 100, breaks = NULL,
  cex.legend = 1, cex.names = 1, cex.values = cex.names, digits = 2L,
  mar.min = 2, add = FALSE, interp = FALSE, grid = 200,
  land.only = FALSE, extrap = FALSE, outside = FALSE, type = "interp",
  linear = FALSE, k = 40, p.pch = 15, p.cex = 1, shift = NULL,
  trans = NULL, ...)

Arguments

map	the map to be plotted, the map object must be a list of matrices with first column indicating the x coordinate and second column the y coordinate each, also see polygon.
x	a matrix or data frame with two columns, first column indicates the region and second column the the values which will define the background colors of the polygons, e.g. fitted values from bayesx. More columns are possible, e.g. quantiles, which can accessed with argument se.
id	if argument x is a vector, argument id should contain a character vector of the same length of x with entries indicating the polygon the i-th value of x belongs to, i.e. id must contain the same names as polygon names in map.
c.select	select the column of the data in x which should be used for plotting, may be an integer or character with the corresponding column name.
legend	if set to TRUE, a legend will be shown.
missing	should polygons be plotted for which no data is available in x?
swap	if set to TRUE, colors will be represented in reverse order.
range	specify the range of values in x which should enter the plot, e.g. only values between -2 and 2 are of interest then range = c(-2, 2).
names	if set to TRUE the name for each polygon will also be plotted at the centroids of the corresponding polygons.
values	if set to TRUE the corresponding values for each polygon will also be plotted at the centroids of the polygons.
col	the color of the surface, may also be a function, e.g. col.surface = heat.colors.
ncol	the number of different colors that should be generated if col is a function.
breaks	a set of breakpoints for the colors: must give one more breakpoint than ncol.
cex.legend	text size of the numbers in the legend.
cex.names	text size of the names if names = TRUE.
cex.values	text size of the names if values = TRUE.
digits	specifies the legend decimal places.
mar.min	Controls the definition of boundaries. Could be either NULL for individual settings of mar or a value which defines mar as follows: The boundaries will be calculated according to the height to width ratio of the map with minimal boundary mar.min.
add	if set to TRUE, the map will be added to an existing plot.
interp	logical. Should the values provided in argument x be interpolated to obtain a smooth colored map.
grid	integer. Defines the number of grid cells to be used for interpolation.
land.only	if set to TRUE, only interpoltated pixels that cover land are drawn, see also function map.where.
extrap	logical. Should interpolations be computed outside the observation area (i.e., extrapolated)?
outside	logical. Should interpolated values outside the boundaries of the map be plotted.
type	character. Which type of interpolation metjod should be used. The default is type = "interp", see function interp. The two other options are type = "mba", which calls function mba.surf of package MBA, or type = "mgcv", which uses a spatial smoother withing package mgcv for interpolation. The last option is definitely the slowest, since a full regression model needs to be estimated.
linear	logical. Should linear interpolation be used withing function interp?
k	integer. The number of basis functions to be used to compute the interpolated surface when type = "mgcv".
p.pch	numeric. The point size of the grid cells when using interpolation.
p.cex	numeric. The size of the grid cell points whein using interpolation.
shift	numeric. Constant to be added to the smooth before plotting.
trans	function to be applied to the smooth before plotting, e.g., to transform the plot to the response scale.
...	parameters to be passed to colorlegend and others, e.g. change the border of the polygons and density, see polygon. Please see the exmaples.

Author(s)

Nikolaus Umlauf, Thomas Kneib, Stefan Lang, Achim Zeileis.

See Also

plot.bayesx, read.bnd, colorlegend.

Examples

## load a sample map
data("FantasyBnd")

## plot the map 
op <- par(no.readonly = TRUE)
plotmap(FantasyBnd, main = "Example of a plain map")
plotmap(FantasyBnd, lwd = 1, main = "Example of a plain map")
plotmap(FantasyBnd, lwd = 1, lty = 2)
plotmap(FantasyBnd, lwd = 1, lty = 2, border = "green3")
plotmap(FantasyBnd, lwd = 1, lty = 2, border = "green3",
  density = 50)
plotmap(FantasyBnd, lwd = 1, lty = 2, 
  border = c("red", "green3"),
  density = c(10, 20), angle = c(5, 45))
plotmap(FantasyBnd, lwd = 1, lty = 2, 
  border = c("red", "green3"),
  density = c(10, 20), angle = c(5, 45),
  col = c("blue", "yellow"))
plotmap(FantasyBnd, col = gray.colors(length(FantasyBnd)))

## add some values to the corresponding polygon areas
## note that the first column in matrix val contains
## the region identification index
x <- cbind(as.integer(names(FantasyBnd)), runif(length(FantasyBnd), -2, 2))
plotmap(FantasyBnd, x = x)
        
## now only plot values for some certain regions
set.seed(432)
samps <- sample(x[,1], 4)
nx <- x[samps,]
plotmap(FantasyBnd, x = nx, density = 20)

## play with legend
plotmap(FantasyBnd, x = x, names = TRUE, legend = FALSE)
plotmap(FantasyBnd, x = nx, density = 20, pos = c(0, 1))
plotmap(FantasyBnd, x = nx, density = 20, pos = c(0, 0.8),
  side.legend = 2)
plotmap(FantasyBnd, x = nx, density = 20, pos = c(0, 0.8),
  side.legend = 2, side.tick = 2)
plotmap(FantasyBnd, x = nx, density = 20, pos = c(0, 0.8),
  side.legend = 2, side.tick = 2, cex.legend = 0.5)
plotmap(FantasyBnd, x = x, values = TRUE, 
  pos = c(-0.15, -0.12))
plotmap(FantasyBnd, x = nx, values = TRUE, 
  pos = c(-0.07, -0.22), width = 2, 
  at = nx[,2], side.legend = 2, distance.labels = 3,
  density = 20)
plotmap(FantasyBnd, x = nx, values = TRUE, 
  pos = c(-0.07, -0.22), width = 2, 
  at = nx[,2], side.legend = 2, distance.labels = 3,
  density = 20, symmetric = FALSE, 
  col = heat_hcl, swap = TRUE)
plotmap(FantasyBnd, x = nx, values = TRUE, 
  pos = c(-0.07, -0.22), width = 2, 
  at = nx[,2], side.legend = 2, distance.labels = 3,
  density = 20, symmetric = FALSE, 
  col = heat_hcl, swap = TRUE, range = c(-5, 5))
plotmap(FantasyBnd, x = nx, values = TRUE, 
  pos = c(-0.07, -0.22), width = 2, 
  at = nx[,2], side.legend = 2, distance.labels = 3,
  density = 20, symmetric = FALSE, 
  col = heat_hcl, swap = TRUE, lrange = c(-5, 5))
plotmap(FantasyBnd, x = nx, values = TRUE, 
  pos = c(-0.07, -0.22), width = 2, 
  at = nx[,2], side.legend = 2, distance.labels = 3,
  density = 20, symmetric = FALSE, 
  col = heat_hcl, swap = TRUE, 
  ncol = 4, breaks = seq(-2, 2, length = 5))

## more position options
plotmap(FantasyBnd, x = nx, density = 20, pos = "bottomleft")
plotmap(FantasyBnd, x = nx, density = 20, pos = "topleft")
plotmap(FantasyBnd, x = nx, density = 20, pos = "topright")
plotmap(FantasyBnd, x = nx, density = 20, pos = "bottomright")
plotmap(FantasyBnd, x = nx, density = 20, pos = "right")
par(op)

# load and plot a map from GermanyBnd
op <- par(no.readonly = TRUE)
data("GermanyBnd")
plotmap(GermanyBnd, main = "Map of GermanyBnd")
n <- length(GermanyBnd)

# add some colors
plotmap(GermanyBnd, col = rainbow(n))
plotmap(GermanyBnd, col = heat.colors(n))
plotmap(GermanyBnd, col = topo.colors(n))
plotmap(GermanyBnd, col = cm.colors(n))
plotmap(GermanyBnd, col = gray.colors(n))
plotmap(GermanyBnd, col = c("green", "green3"))
par(op)


## now with bayesx
set.seed(333)

## simulate some geographical data
data("MunichBnd")
N <- length(MunichBnd); names(MunichBnd) <- 1:N
n <- N*5

## regressors
dat <- data.frame(id = rep(1:N, n/N))
dat$sp <- with(dat, sort(runif(N, -2, 2), decreasing = TRUE)[id])

## response
dat$y <- with(dat, 1.5 + sp + rnorm(n, sd = 0.6))

## Not run: 
## estimate model
b <- bayesx(y ~ sx(id, bs = "mrf", map = MunichBnd),
  method = "MCMC", data = dat)

## summary statistics
summary(b)

## plot spatial effect
op <- par(no.readonly = TRUE)
plot(b, map = MunichBnd)
plot(b, map = MunichBnd, c.select = "97.5
plot(b, map = MunichBnd, c.select = "2.5
plot(b, map = MunichBnd, c.select = "50
plot(b, map = MunichBnd, names = TRUE, 
  cex.names = 0.5, cex.legend = 0.8)
plot(b, map = MunichBnd, range = c(-0.5, 0.5))
plot(b, map = MunichBnd, range = c(-5, 5))
plot(b, map = MunichBnd, col = heat_hcl, 
  swap = TRUE, symmetric = FALSE)
par(op)

## End(Not run)

R2BayesX documentation built on Oct. 20, 2023, 9:11 a.m.