Skip to content

Guerry data: making a package

3 messages · Michael Friendly, Roger Bivand

Original
Michael Friendly
# 
Hi

Several years ago I sought & received help (thanks!) from people on this 
list regarding map and data
files associated with my article, A.M. Guerry's Moral Statistics of 
France: Challenges for multivariate spatial analysis,
that appeared in Statistical Science, 2007, 22,368-399.  As 
supplementary materials for that article, I prepared
http://www.math.yorku.ca/SCS/Gallery/guerry/

and also included map and data files for R at
http://www.math.yorku.ca/SCS/Gallery/guerry/maps.html#R
but these were really just a sketch that required more work to make them 
fully usable.

Now I'd like to make an R package including the gfrance1 map and the 
guerry data, and other things,
but I found that something in maptools or other packages had changed, so 
that a script that worked
for me with these map files back in Aug, 2008 does so no longer:

 > library(sp)
 > library(maptools)
 > gfrance <- readShapeSpatial("gfrance1")
Error in read.shape(filen = fn, verbose = verbose, repair = repair) :
  File size and implied file size differ, consider trying repair=TRUE

I get the same error with readShapePoly().

Once I can get that sorted out, it's not exactly clear to me exactly how 
to prepare map files
(SpatialPolygonsDataFrame) for a package with appropriate 
documentation.  Using
another version of the Guerry map and data (omitting Corsica) provided 
by Stephane Dray, I *could* read it
successfully, then tried

 > cd("dray")
 > gfrance85 <- readShapePoly("gfrance1")
 > save(gfrance85, file="gfrance85.rda")
 > prompt(gfrance85)
Created file named 'gfrance85.Rd'.
Edit the file and move it to the appropriate directory.

There is no prompt method for SpatialPolygonsDataFrame objects, so 
prompt() just
dumps the entire results of str(gfrance85); I can of course edit out the 
@ polygons list.
Is this the recommended way to include such objects in packages? 

Thanks,
-Michael
12 days later
Roger Bivand
#
On Sat, 17 Oct 2009, Michael Friendly wrote:

            

      
      
      
    

        
readShapeSpatial() is a wrapper for readShapePoly() and others, choosing 
on shape type. The message suggests that the *.shx file is not as 
expected.

With a freshly downloaded copy from your website, I cannot reproduce the 
problem, and gfrance1.shp reads correctly on:

            

      
      
      
    

        
R version 2.10.0 (2009-10-26)
i686-pc-linux-gnu

locale:
  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
  [5] LC_MONETARY=C              LC_MESSAGES=en_US.UTF-8
  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C
  [9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] maptools_0.7-26 foreign_0.8-38  sp_0.9-44

loaded via a namespace (and not attached):
[1] grid_2.10.0     lattice_0.17-26

            

      
      
      
    

        
The prompt method for data either identifies data.frame objects or dumps 
the object. For me,

prompt(as(gfrance, "data.frame"), name="gfrance")

looks easier to edit, adding the class details.

Hope this helps,

Roger

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Michael Friendly
    

    

      
      #
    

  


  

      
Thanks, Roger

I was able to sort out my problems with reading the Guerry data/map with 
the help of Stephane Dray.
There is now an R-Forge project, Guerry, containing what we have 
assembled so far.
https://r-forge.r-project.org/projects/guerry/

In the .Rd files for the SpatialPolygonsDataFrames, gfrance & gfrance85 
included in the package, I
note that no PROJ4 information is provided, because we weren't sure 
exactly what to specify.

On my web page, I noted that you had earlier suggested
The projection here seems to be France ED50 EuroLambert, Lambert Conic 
Conformal with parameters set for continental France. Roger says: In 
PROJ.4 this is:

+proj=lcc +lat_1=46.8 +lat_0=46.8
+lon_0=2.337229166666667 +k_0=0.99987742 +x_0=600000 +y_0=2200000
+ellps=intl +units=m 

Given that:

 > slot(gfrance,"proj4string")
CRS arguments: NA

what can I specify to make this more spatial-friendly in a package?  I 
don't know what a lot of
that stuff means, but if I'm creating a package, I'd like it to be 
minimally sufficient for others to
use in different contexts.

-Michael

        
Roger Bivand wrote: