[R] Maps in R

[Ross Ihaka]

On Sun, 18 Apr 1999, Tim Churches wrote:

> Apologies in advance from an R newbie if this is a dumb question. Is it
> possible to produce choropleth maps in R? I gather from some nice research
> papers on the AT&T (sorry, Lucent) S web site that in S (and S Plus) there
> is a map() function which uses arrays of polylines and regions to draw map
> borders and then fill them with patterns or shades according to some scalar
> quantity. We currently produce such maps (of disease rates) using PROC GMAP
> in SAS/GRAPH which also uses the same sort of polyline format for definition
> of the areas. The problem in SAS is that although the colours and sharding
> can be precisely controlled, the means of allocating those colours to
> regions on the map or ranges in the data stinks.
> 
> If the map() function hasn't been implemented in R, are there any plans?

I am most of the way through porting the Becker/Wilks code to R and I will
be finishing this some time in the next few months.  Fritz Leisch has also
expressed an interest and I passed what I had along to him. The remaining
problems to solve are:

  1. Thinning the polygons which make up the map (so they will print in
      postscript).

  2. Assembling a reasonable collection of map projections.

There is nothing difficult in this, its just a matter of finding a few
days to do the job.

In addition I'm looking at putting a better set of maps together (I'm
expecting the US National Imagery and Mapping Agency 4 CD set to show up
on my doorstep at any minute).

If you have Java expertise available, you might also want to have a look
at Openmap from BBN <http://openmap.bbn.com>.  It looks like a pretty good
framework for doing high-quality mapping work. Now if we just had that
R/Java link which John Chambers and Duncan Temple Lang are working on ...

> I must say that we have been mightily impressed by the scope and stability
> of R. We are using it under Win NT at present but we are about to set it up
> on a Linux box. I am very much hoping that we can spend this year's
> discretionary budget on a nice new and fast Linux box (stuffed with RAM) for
> running R rather than spending the same amount on two S Plus licenses for
> Windows (which are not exactly cheap since we are a government department,
> although since we started paying for SAS each year nothing shocks us any
> more...).

I know the feeling, though we fare a little better because we are
educational.  My department is however looking into getting SAS Data
Enterprise Miner -- which should effectively absorb any amount of budget
excess we might have.

	Ross

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._