[R] from long/lat to UTM

[Greg Snow]

>>> Ted Harding <Ted.Harding at nessie.mcc.ac.uk> 03/10/05 09:13AM >>>
>> On 10-Mar-05 Gabor Grothendieck wrote:
>> >  <Ted.Harding <at> nessie.mcc.ac.uk> writes:
>> >: [...]
>> >: As Sander says, "Not sure where you would derive the time
zone!".
>> >: 

[snip]

>> The only completely general mechanism I can think of would
>> consist of
>> 
>> a) a named list of boundary contours

I found a shapefile of time zones at: 
http://openmap.bbn.com/data/shape/timezone/ 

A google search on the keywords timezone & shapefile finds a 
few other possibilities. 

>> b) a function which, for each name in the list, returns
>>    the TZ name

The "maptools" package can read in the above shapefile and 
convert it to a list of polygons and the "sgeostat" has an
"in.polygon" function to see if a point is in a given polygon.

The following code worked for me:

library(maptools)
tz <- read.shape('c:/maps/WrldTZA')
plot(tz)
plot(tz,xlim=c(-150,-50), ylim=c(20,50))

mappoly <- Map2poly(tz)

library(sgeostat)

tmp <- sapply(mappoly, function(x){ x <- na.exclude(x)
	in.polygon( -110, 42, x[,1], x[,2] ) } )

tz$att.data[tmp,]


the -110 and 42 come from me plotting the map and using 
locator to approximate where I am at.  I was actually 
suprized at how quick the sapply was (under a second on
my fairly fast pc (windows 2000)).

It shouldn't be too hard to convert this into a more general
function.




Greg Snow, Ph.D.
Statistical Data Center
greg.snow at ihc.com
(801) 408-8111