Skip to content

Linking R to GIS

2 messages · Gatehouse, Simon (DEM, North Ryde), Roger Bivand

#
Dear All,
Has anyone worked with linking R to both ArcView and MapInfo GIS systems? My
desire is to capture object selection from GIS and pass to R and to return
spatially based statistical model results from R to GIS. If both could
simultaneously access the same data file this would be a bonus.

My inclination is to put together some routines based on ODBC linking but
thought I would seek like minded people before 
pressing too far into the evaluation.  Any thoughts or suggested alternative
directions or shared examination would be welcome.

SimonG

****************************************************************************
*****************************
Simon Gatehouse                                  Phone: 61 (2) 9490 8677
Geochemist                                           Mobile:61 0407 130 635
CSIRO Exploration & Mining                   Fax:   61 (2) 9490 8921
 
Address:
CSIRO Division of Exploration and Mining,
Newbigin Close off Julius Ave
North Ryde, NSW 2113
 
Mail:     
CSIRO Division of Exploration and Mining,
P.O. Box 136
North Ryde, NSW  1670
Australia
 
E-mail:          simon.gatehouse at dem.csiro.au
Web Page:        http://www.csiro.au/
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
#
On Tue, 26 Jun 2001, Gatehouse, Simon (DEM, North Ryde) wrote:

            
One general comment and a couple of suggestions. Both ArcView and MapInfo
treat the geometry of spatial objects in a simplifying way, not handling
topology in the thorough way that heavy-duty GIS - like Arc-Info - do.
While most scientists are afforded access to desk-top style GIS such as
these, there are valid issues to be addressed about the need for topology
in interfacing (not least unambiguity in choice of spatial objects). The S
map.builder (http://lib.stat.cmu.edu/general/map.builder) does construct
the topology, GRASS does too, both for legacy vector and the forthcoming
5.1 beta vector revision.

Not forgetting that, your idea of doing the selection in the GIS and
getting R to interface the attribute table (ie. doing aspatial analysis on
the attributes in R) via ODBC is potentially viable. It might be feasible
to prototype it using the connections interface family now in R-base - see
the R Data Import/Export manual. Even writing a *.csv file might do - or
maybe the DCOM package under contrib/extra/dcom at CRAN. However, this
just gets the attribute data into R and back again.

I'm sure there are plenty of people with similar problems who would be
interested in how you progress.

Roger