moranbi.cluster {bispdep}

Maps of BiLISA clusters and statistical significance associated with BiLISA

Description

Using the localmoran.bi function, create a Bivariate Local Indicators of Spatial Association (BiLISA) cluster map and corresponding significance map. The maps are made by calculating the BiLISAs from localmoran.bi for each of the spatial units and then the statistical significance is evaluated to determine the spatial clusters and outliers.

Usage

moranbi.cluster(varY, varX, listw, zero.policy = NULL, polygons, conditional=TRUE,
                significant = TRUE, alternative = "two.sided", pleg, ...)

Arguments

varY

a numeric vector the same length as the neighbours list in listw and x, with the values of the secundary variable y

varX

a numeric vector the same length as the neighbours list in listw and y, with the values of the primary variable x

listw

a neighbours list with spatial weights. From package spdep: a listw object. Use poly2nb (class nb) and nb2listw (class listw, nb) from package spdep. Can be any type of listw object, for instance, rook contiguity (common edge) or queen contiguity (common edge or common vertex)

zero.policy

by default = NULL, if FALSE stop with error for any empty neighbour sets, if TRUE permit the weights list to be formed with zero-length weights vectors. Parameter inherited from the spdep package.

polygons

SpatialPolygons, SpatialPolygonsDataFrame or sfc_POLYGON object

conditional

default TRUE: expectation and variance are calculated using the conditional randomization null (Sokal 1998 Eqs. A7 & A8). Elaboration of these changes available in Sauer et al. (2021). If FALSE: expectation and variance are calculated using the total randomization null (Sokal 1998 Eqs. A3 & A4).

significant

by default is TRUE, if FALSE the significant map is not created

alternative

by default is "two.sided". Type of alternative hypothesis test. Other values are "less" or "greater".

pleg

the x and y co-ordinates to be used to position the legend. They can be specified by keyword or in any way which is accepted by xy.coords

...

other parameters similar to internal function moran and other graphical parameters as in par(..)

Details

Using the function localmoran.bi create the Bivariate Local Indicators of Spatial Association - BiLISA Cluster Map and the corresponding Significance Map. The significance map is done testing the null hypothesis (Ho) of zero spatial autocorrelation for each spatial unit and then plotting a choropleth map with this legend values: (Not Significant, p-value=0.05, p-value= 0.01, p-value=0.001, p-value=0.0001, and Neighborless). Maps can represent concentrations of similar (cluster) or dissimilar values (spatial outliers). Most significant clustered spatial units are those with p-values smaller than 0.0001. Not significant clustered spatial units are those with p-values greater than 0.05. BiLISA Cluster Map is done based on the significance map but the choropleth legend is different (Not - Significant, High-High, Low-Low, Low-High, High-Low, and Neighborless).

Value

one or two maps

Links

1. Indicators of Spatial Association

See Also

• Bivariate Moran's Ixy: moran.bi

• Plot Bivariate Moran's Ixy: moranbi.plot

• Bivariate Moran's Ixy Test: moranbi.test

• Create object "nb": poly2nb

• Create object "listw"/"nb": nb2listw

Examples

library(spdep)
data(columbus)
columbus <- st_read(system.file("shapes/columbus.shp", package="spData")[1], quiet=TRUE)
plot(st_geometry(columbus))
col_nbq <- poly2nb(columbus)
a.lw <- nb2listw(col_nbq, style="W")
moranbi.cluster(columbus$CRIME, columbus$HOVAL, a.lw, zero.policy = FALSE,
                conditional=TRUE, st_geometry(columbus), significant=TRUE,
                pleg = "topleft")
moranbi.cluster(columbus$CRIME, columbus$HOVAL, a.lw, zero.policy = FALSE,
                st_geometry(columbus), significant=TRUE, alternative="greater",
                pleg = "topleft")