Areal or lattice data is observed on, or often aggregated up to often arbitraty spatial units, like census tracts, counties, or countries. Often, such data is analysed as though the observations were independent, ignoring possible spatial dependence. In practice, dependence may stem from the misfitting of unit boundaries to phenomena being measured (labour markets may not match city limits), from missing variables resolving dependence, or from the actual impact of neighbours on each other (spillover).

Spatial data analysis
R spatial projects

Packages on CRAN
Maps
Point pattern
Geostatistics
Areal
GIS interfaces
Others

Other Packages
Details

Mailing list
R-sig-Geo

Links
Back to SAL

spdep

spdep is a collection of functions to create spatial weights matrix objects from polygon contiguities, from point patterns by distance and tesselations, for summarising these objects, and for permitting their use in spatial data analysis; a collection of tests for spatial autocorrelation, including global Moran's I, Geary's C, Hubert/Mantel general cross product statistic, Empirical Bayes estimates and spatial Index, and Getis/Ord G, local Moran's I and Getis/Ord G, saddlepoint approximations for global and local Moran's I; and functions for estimating spatial simultaneous autoregressive (SAR) models. It now also provides weighted SAR and CAR fitting functions, and Spatial Filtering.

It has no homepage as such, but its CVS repository is on Sourceforge here.

Releases in the 0.3-* series depend on the SparseM package for calculating the determinant of the Jacobian, rather than using the "sparse" code from Netlib (originally from UCB). A number of other bugs have been and continue to be fixed thanks to very useful bug reports posted by users.

Releases from 0.2-3 and onwards depend on the maptools package version 0.3 or higher; they implement the R NAMESPACE mechanism, and this formalises the relationship between the two packages. Install maptools first, then spdep. spdep now also has a "vignette" describing how to use maptools and spdep together with the North Carolina SIDS data set - you can see where the "vignette" files are on your system after installation at the foot of:

help(package=spdep)

DCluster

DCluster is an R package for the detection of spatial clusters of diseases. A description of the package and examples can be found in this paper presented at DSC 2003. It has a homepage.

spgwr

spgwr is a package for Geographically Weighted Regression, implementing some of the methods described in A. Stewart Fotheringham, Chris Brunsdon and Martin Charlton, Geographically Weighted Regression: the analysis of spatially varying relationships, 2002, John Wiley. It depends in sp classes. From release 0.4, it is possible to use snow clusters with some functions in the package.