Skip to content

spatial clusters

5 messages · dorina.lazar, Georg Ruß, Dorina Lazar +2 more

Original
dorina.lazar
# 
Dear all,

I try to use SKATER algorithm (from package spdep) to classify the
countries in clusters taking account of the geographical position and from
macroeconomic indicators.

How to create a shapefile including my data about some socio-demographic
indicators (7),  for about 30 countries around the world (to be read with
readShapePoly)? 

Would be useful to suggest some good introductory lectures in spatial
statistics (links or books).  

Thanks you for your help.
Dorina


PS. Roger, Georg,  thanks a lot for your suggestions.
Georg Ruß
#
On 16/12/10 09:06:33, dorina.lazar wrote:
Hi Dorina,

I'm not quite sure what you have and what you want.

I assume that you'd like to create a shapefile such that for every country
(which is represented as a polygon) there's a vector of seven attributes
attached to it:

country1:  attr1 attr2 attr3 attr4 attr5 attr6 attr7
country2:  attr1 attr2 attr3 attr4 attr5 attr6 attr7
...

Do you have your countries as spatial polygons in R (SpatialPolygons-
DataFrame or similar)? If yes, you should be able to add your socio-demo-
graphic indicators to this data frame (just as in normal data frames). In
this way you'd join the data you have inside R.

If you have the shapefile for the countries available and would like to
edit this directly to add the indicators, I'd suggest using a GIS like
GRASS http://grass.fbk.eu/ Afterwards, this can be easily read into R.
I think the following is what Roger would suggest:

 author = {Bivand, Roger S. and Pebesma, Edzer J. and G?mez-Rubio, Virgilio},
 title = {Applied Spatial Data Analysis with R},
 series = {Use R},
 publisher = {Springer},

http://www.asdar-book.org/
http://www.springerlink.com/content/uw07v1/
(Depending on your login/institute you may have fulltext access, but the
book is definitely worth buying anyway.)

Regards,
Georg.
--
Research Assistant
Otto-von-Guericke-Universit?t Magdeburg
research at georgruss.de
http://research.georgruss.de
2 days later
Dorina Lazar
# 
Georg,

Thanks for being very patient with me. 
 
I have not a shapefile for countries.

I have tried to obtain it. I have loaded a shapefile for Finland, GIS data, from  http://www.eea.europa.eu/data-and-maps/data/eea-reference-grids/zipped-shape-file-finland

and I tried to read the file FI_1K.shp
Fin<- readShapePoly(system.file("etc/shapes/ FI_1K.shp", package="spdep"))
but...
Error in getinfo.shape(filen) : Error opening SHP file

How can I obtain, for each country,  the shapefile spatial polygon in R (SpatialPolygons-DataFrame)?

Thanks again,
Dorina
--- On Thu, 12/16/10, Georg Ru? <research at georgruss.de> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        
      

    

    

      
      

        


  

        

      Roger Bivand
    

    

      
      #
    

  


  

      
On Sat, 18 Dec 2010, Roman Lu?trik wrote:

        

            

      
      
      
    

        
I don't think that we know which countries Dorina needs, or at which 
administrative level. If there are only 30 country observations, and the 
countries are not contiguous, then the whole exercise seems unnecessary. 
Can Dorina use the simplified world map in the maptools package:

library(maptools)
?wrld_simpl

adding the data using the ISO country names? If these are countries at 
some administrative division level, then maybe www.gadm.org will have what 
is needed, but the countries will have to be joined together, or the whole 
map subsetted. If these are say EU NUTS2 regions, then use can be made of 
GISCO shapefiles from Eurostat. But we don't know what boundaries are 
needed, and of course to use Skater for clustering, you need the 
boundaries.

Roger