[R-sig-Geo] Rendering a map of all zip codes (ZCTAs)

[Roger Bivand]

On Tue, 18 Feb 2014, arilamstein at gmail.com wrote:

> I recenty posted<https://groups.google.com/forum/#!topic/ggplot2/5ZTGjl_Guvo)>this
> question on the ggplot2 mailing list and did not hear back, so I
> thought that reposting to sig-geo might be better.
>
> I have recently released an R package called
> choroplethr<http://tech.truliablog.com/2014/01/15/the-choroplethr-package-for-r/>that
> attempts to facilitate the creation of choropleth maps in R at the
> state, county and ZIP code level of detail.  It uses the state and county
> maps that come with the ggplot2 package (or is the it the maps package?).
> For zip maps it just renders a scatterplot against the state boundaries.

Your "blog" points to github - the package is on CRAN, point to that; 
install.github() tries to install from source, and most users need binary 
installs; it is also not >= GPL-2. Note that the boundary data in the maps 
package do not have clear pedigrees, and the zipcode package only ships 
points (so why would you be surprised if only points are displayed?) - 
they are also without clear authority.

>
> I would like to render zip code choropleths using the ZIP Code Tabulated
> Area (ZCTA) shapefile that the census distributes
> here<http://www.census.gov/cgi-bin/geo/shapefiles2013/main>.
> Hadley has a nice tutorial on how to work with shapefiles in ggplot2
> here<https://github.com/hadley/ggplot2/wiki/plotting-polygon-shapefiles>.
> However, those directions do not work for me, and I think that it might be
> due to the size of the shapefile (501 MB zipped).  As I mentioned on the
> ggplot2 mailing list, the call to "fortify" will run all night and still
> not finish.

You can choose to use ggplot2 and fortify, but you go from fish to fish 
soup without the possibility of getting the fish back (in reasonable 
time, fish paste best option). However, converting all the coordinates to 
data.frame format is time-consuming, and creates very many copies of 
attrribute data.

>
> Can anyone give me pointers on how to work with this file effectively in R?
> Ideally I would be able to distribute this map as part of my package as
> well, but clearly the current size of the shapefile makes that prohibitive.
>

Firstly, do study thematic cartography. You'll find that many of the 30K 
geometries are invisible on most available screen resolutions, so that 
visualising more than a small subset isn't very sensible anyway (single 
pixels belonging to different class intervals from their neighbours are 
very hard to pick out, even with transparent borders). So choosing the 
subset of your imported SpatialPolygonsDataFrame object you need and only 
fortifying that subset will make things more tractable.

You should drop the idea of shipping the US Census shapefile of ZCTA, CRAN 
and its mirrors do not need to duplicate files that are available online 
elsewhere; if the boundaries were on a WFS (Web Feature Server), it might 
be possible to download just the subset needed.

Roger

> Thanks.
>
> Ari
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no