predictive modeling for crime rates in Chicago wards
Definitions for "community areas" in Chicago can be downloaded as GIS shapefiles.
The shapefile file data/chicomm.shp comes from http://www.lib.uchicago.edu/e/collections/maps/chicomm.zip
Some discussion of how to convert from shapefiles to more general purpose polygons is given in http://www.quora.com/How-can-I-convert-a-shapefile-to-a-lat-long-polygon-array-for-use-in-Google-Maps
Download, build, and install the GDAL and OGR utilities from http://www.gdal.org/ogr2ogr.html
./configure ; make; sudo make install
Download, build, and install the GEOS utilities from http://download.osgeo.org/
./configure ; make; sudo make install
Install the lxml package from http://lxml.de/ and the shapely package from https://pypi.python.org/pypi/Shapely
based on instructions in http://osxastrotricks.wordpress.com/2010/07/19/install-python-shapely/
Next, we convert the shapefile to KML, then parse the KML to construct Shapely polygons as "LinearRings", based on http://toblerity.github.io/shapely/manual.html#linearrings
ogr2ogr -f "KML" foo.kml data/chicomm.shp
./src/py/shape.py foo.kml