Making a choropleth¶

Choropleths are a way to represent some non-geographical data that varies by regions on your map. We'll be making choropleths with Folium and Geopandas.

We're going to start with a dataset from the World Bank on Internet usage in 2016.

In [1]:

import pandas
import matplotlib.pyplot as plt

In [2]:

internet = pandas.read_csv('data/world_internet_usage_2016.csv', skiprows=18, names=["name", "internet_use"])
internet.head()

Out[2]:

	name	internet_use
17	Afghanistan	10.6
18	Albania	66.4
19	Algeria	42.9
20	American Samoa	NaN
21	Andorra	97.9

In [3]:

internet[internet.internet_use.isnull()]

Out[3]:

	name	internet_use
20	American Samoa	NaN
44	British Virgin Islands	NaN
56	Channel Islands	NaN
110	Isle of Man	NaN
119	Korea, Dem. People’s Rep.	NaN
121	Kosovo	NaN
154	Nauru	NaN
162	Northern Mariana Islands	NaN
166	Palau	NaN
180	San Marino	NaN
188	Sint Maarten (Dutch part)	NaN
199	St. Martin (French part)	NaN
217	Turks and Caicos Islands	NaN

In [4]:

internet.dropna(inplace=True)
internet.head()

Out[4]:

	name	internet_use
17	Afghanistan	10.6
18	Albania	66.4
19	Algeria	42.9
21	Andorra	97.9
22	Angola	13.0

Geopandas¶

In [5]:

import geopandas
world = geopandas.read_file(geopandas.datasets.get_path('naturalearth_lowres'))
world.head()

Out[5]:

	pop_est	continent	name	iso_a3	gdp_md_est	geometry
0	28400000.0	Asia	Afghanistan	AFG	22270.0	POLYGON ((61.21081709172574 35.65007233330923,...
1	12799293.0	Africa	Angola	AGO	110300.0	(POLYGON ((16.32652835456705 -5.87747039146621...
2	3639453.0	Europe	Albania	ALB	21810.0	POLYGON ((20.59024743010491 41.85540416113361,...
3	4798491.0	Asia	United Arab Emirates	ARE	184300.0	POLYGON ((51.57951867046327 24.24549713795111,...
4	40913584.0	South America	Argentina	ARG	573900.0	(POLYGON ((-65.50000000000003 -55.199999999999...

In [6]:

len(set(world.name) & set(internet.name))

Out[6]:

In [7]:

len(set(world.name) - set(internet.name))

Out[7]:

In [8]:

len(set(internet.name) - set(world.name))

Out[8]:

In [9]:

world = world.merge(internet, how='left', on='name')
world.head()

Out[9]:

	pop_est	continent	name	iso_a3	gdp_md_est	geometry	internet_use
0	28400000.0	Asia	Afghanistan	AFG	22270.0	POLYGON ((61.21081709172574 35.65007233330923,...	10.6
1	12799293.0	Africa	Angola	AGO	110300.0	(POLYGON ((16.32652835456705 -5.87747039146621...	13.0
2	3639453.0	Europe	Albania	ALB	21810.0	POLYGON ((20.59024743010491 41.85540416113361,...	66.4
3	4798491.0	Asia	United Arab Emirates	ARE	184300.0	POLYGON ((51.57951867046327 24.24549713795111,...	90.6
4	40913584.0	South America	Argentina	ARG	573900.0	(POLYGON ((-65.50000000000003 -55.199999999999...	70.2

I'm going to copy over some Proj4 strings, that we can use to tell the underlying projection libraries, PyProj and Proj.4, how to project our map.

In [10]:

mercator = '+proj=merc'
robinson = '+proj=robin'
wgs84 = '+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs'

Next, we'll clean up the map a bit by removing Antarctica (makes projecting a bit more difficult, and is irrelevant to what we're making anyway) and any filling any missing data about Internet connectivity with a zero.

In [11]:

world = world[(world.name!="Antarctica")].fillna(0)

We'll make a copy of the DataFrame when we do the projections, so we don't get any weird geometric artifacts.

In [12]:

# mercator projection
worldMercator = world.to_crs(mercator)
worldMercator.plot(figsize=(20,20), column='internet_use', cmap="OrRd", scheme="QUANTILES", k=5, legend=True)
plt.axis('off')
plt.title("Internet usage by country")
plt.savefig('data/internet_choropleth_geopandas.png')
plt.show()

Data munging¶

...

PySAL¶

...

Folium¶

...

References¶

Folium choropleth: http://python-visualization.github.io/folium/quickstart.html#Choropleth-maps

World bank data: https://data.worldbank.org/data-catalog/country-profiles