#!/usr/bin/env python
# coding: utf-8

# # Plotting with Viridis in PySAL's `geoplot`
# 
# * [Dani Arribas-Bel](http://darribas.org) [[@darribas](http://twitter.com/darribas)]
# 
# This document shows how one can leverage the recent release of [`palettable`](https://github.com/jiffyclub/palettable/releases/tag/v3.0.0) to make choropleths using the [ever famous Viridis](https://www.youtube.com/watch?v=xAoljeRJ3lU) palette.

# In[1]:


get_ipython().run_line_magic('matplotlib', 'inline')

import bokeh as bk
import palettable as pltt
import geopandas as gpd
import pysal as ps
from pysal.contrib import pdio
from pysal.contrib.geotable.utils import to_df, to_gdf
from pysal.contrib.viz.mapping import geoplot

pl = pltt.matplotlib.Viridis_10

bk.plotting.output_notebook()


# Let's load up the US counties, and we will use homicide rates in 1990 (`HR90`) as an example. As a bonus, this is also an example about how you can read data in `geopandas`, reproject it, and push it over to `PySAL`'s new `geotable`!

# In[2]:


gdf = gpd.read_file(ps.examples.get_path('NAT.shp'))
gdf.crs = {'init' :'epsg:4326'}
gdf = gdf.to_crs(epsg=3083)
gdb = to_df(gdf)


# Good to go! Thanks to `geoplot`, this means you can plot with equal ease using both backends to `matplotlib` and `bokeh`.

# * `matplotlib`

# In[3]:


geoplot(gdb, 'HR90', palette=pl, k=10)


# * `bokeh`

# In[4]:


geoplot(gdb, 'HR90', palette=pl, k=10, backend='bk')