cartogram: Fit a diffusion cartogram

In omegahat/Rcartogram: Interface to Mark Newman's cartogram software

Description

This uses Mark Newman's C code to fit the diffusion cartogram developed by Michael Gastner and Mark Newman.

One calls this with a matrix (or a data.frame) containing the counts or densities (e.g. population, number of incidences of a particular disease, etc.) for that location. By default, the code returns a list of two matrices (x and y) each with one more row and column as there are in pop. These two "grids" can then be used to predict or compute the transformed or projected value of a point in the new cartogram using the predict method in this package.

This is a reasonably straightforward interface to Mark Newman's code. Therefore to understand what it does and how to use, read the documentation at http://www-personal.umich.edu/~mejn/cart/doc/, specifically "Using the cart program".

Note that the use of the word "cart" has nothing to do with classification and regression trees.

Usage

cartogram(pop, zero = TRUE, blur = 0, sea = NA)

Arguments

pop	a matrix. This is oriented as Xs along the columns and Y's going down each column. In other words, rows in the R matrix correspond to the same y element in each of the X vectors.
zero	a logical value that indicates whether to expand the grid by 1 more row and column. This is what is done by the cart application by Mark Newman so we repeat it here to allow for obtaining comparable results.
blur	a non-negative real value that is used in the C code to add Gaussian noise to the diffusion.
sea	either a missing value (NA) or a multiplier that is passed to addBoundary to embed the pop matrix within a larger grid. The values of the outer grid are constant and the mean of the original pop matrix is used.

Details

See the paper in the references.

Value

A list with two entries, both matrices with as many rows and columns as pop or the enlarged grid if sea is not an NA.

Author(s)

Duncan Temple Lang created the interface for R

References

"Diffusion-based method for producing density equalizing maps", Michael T. Gastner and M. E. J. Newman, Proc. Natl. Acad. Sci. USA 101, 7499-7504 (2004) http://www.pnas.org/cgi/content/abstract/101/20/7499

See Also

addBoundary

Examples

  filename = system.file("extdata", "uspop.tab.gz", package = "Rcartogram")
  pop = read.table(filename)
  grid = cartogram(pop)

  predict(grid, 3.1, 4.5)

  predict(grid, c(3.1, 2), c(4.5, 4))
 
  # note that this is equivalent to using the interp application in
  # the cart(ogram) distribution with inputs as 2.1 3.4
  # i.e. echo "2.1 3.4" | interp 1024 512 output.dat
  #        

    # now predict lots of values.
  ans <- predict(grid, 
                 runif(10000, 1, 1024), 
                 runif(10000, 1, 512))

omegahat/Rcartogram documentation built on May 24, 2019, 1:55 p.m.