Requires:

• Numpy

• GeoPandas

• Matplotlib

One-shot functionnality

>>> result = quick_stewart('nuts3_data.geojson',
                           "pop1999",
                           span=65000,
                           beta=3,
                           resolution=48000,
                           mask='nuts3_data.geojson',
                           nb_class=10,
                           user_defined_breaks=None,
                           output="geojson")

Object-oriented API, allowing to easily redraw contours with new breaks values

>>> StePot = SmoothStewart('nuts3_data.geojson', "pop1999",
                           span=65000, beta=3,
                           resolution=60000,
                           mask='nuts3_data.geojson')
>>> res = StePot.render(nb_class=8, disc_func="jenks",
                        output="GeoDataFrame")
>>> res.plot(cmap="YlOrRd", linewidth=0.1)

The long part of the computation is done during the initialization of SmoothStewart instance (i.e. actually computing potentials). Some convenience methods allows to tweak and re-export the few last steps :

Allow to quickly redraw polygons with a new classification method

Availables classification methods are: “equal_interval”, “prog_geom”, “jenks”, “percentiles” and “head-tail-breaks”

>>> res = StePot.render(nb_class=6,
                        disc_func="percentiles",
                        output="GeoDataFrame")

Allow to set custom break values (highly recommended after a first rendering or having take a look at the distibution):

>>> my_breaks = [0, 1697631, 3395263, 5092894, 6790526,
                 8488157, 10185789, 11883420, 13581052]

>>> res = StePot.render(nb_class=6, user_defined_breaks=my_breaks,
                        output="GeoDataFrame")

From PyPI :

$ pip install smoomapy

From github :

$ git clone http://github.com/mthh/smoomapy.git
$ cd smoomapy/
$ python setup.py install