Non contiguous cartogram in Python

Cartograms are a kind a map that changes region size and/or shape according to a numerical variable. It's for example used to highlight population density differences between countries.

This post explains how to create a non-contiguous cartogram, which is a kind of cartogram that preserves shape but modify sizes. We'll go over a concrete example to illustrate difference in population density in Asia.

About cartograms

To give you a visual idea, here is the cartogram we will step-by-step create in this post:

preview final chart

Libraries & Data

For creating this chart, we will need to load the following libraries:

matplotlib for plotting the chart
geopandas and geoplot: for spatial data plotting
pandas for loading the data
pypalettes: for the color palette
highlight_text for the annotations

# matplotlib tools
import matplotlib.pyplot as plt
from matplotlib.font_manager import FontProperties

# map libraries
import geopandas as gpd
import geoplot as gplt
import geoplot.crs as gcrs

# colors
from pypalettes import load_cmap

# annotations
from highlight_text import fig_text, ax_text

# data manipulation
import pandas as pd

# increase resolution
plt.rcParams['figure.dpi'] = 300
plt.rcParams['savefig.dpi'] = 300

Dataset

Let's start by loading shape data:

world = gpd.read_file('https://raw.githubusercontent.com/holtzy/The-Python-Graph-Gallery/master/static/data/all_world.geojson')
world.head()

	name	geometry
0	Fiji	MULTIPOLYGON (((180.00000 -16.06713, 180.00000...
1	Tanzania	POLYGON ((33.90371 -0.95000, 34.07262 -1.05982...
2	W. Sahara	POLYGON ((-8.66559 27.65643, -8.66512 27.58948...
3	Canada	MULTIPOLYGON (((-122.84000 49.00000, -122.9742...
4	United States of America	MULTIPOLYGON (((-122.84000 49.00000, -120.0000...

Then we load data about the Asian population and surfaces

# get asian population dataset
url = 'https://raw.githubusercontent.com/holtzy/The-Python-Graph-Gallery/master/static/data/asia.csv'
asia = pd.read_csv(url)
asia.head()

	Country	Total Population	Surface Area (sq. km)
0	Russia	1.444444e+08	17098250.0
1	China	1.425671e+09	9600013.0
2	India	1.428628e+09	3287259.0
3	Kazakhstan	1.960663e+07	2724902.0
4	Saudi Arabia	3.694702e+07	2149690.0

Once we have our 2 datasets, we can merge them and create pop_norm_surface column as a measure of population density:

# merge the datasets together
data = world.merge(asia, how='right', left_on='name', right_on='Country')

# filter the data
data = data[['Country', 'geometry', 'Total Population', 'Surface Area (sq. km)']]
data = data[~data['Country'].isin(['Russia', 'Bangladesh', 'Lebanon'])]
data.dropna(inplace=True)
data['pop_norm_surface'] = data['Total Population'] / data['Surface Area (sq. km)']

# display first rows
data.columns = ['Country', 'geometry', 'pop', 'surfaces', 'pop_norm_surface']
data.head()

	Country	geometry	pop	surfaces	pop_norm_surface
1	China	MULTIPOLYGON (((109.47521 18.19770, 108.65521 ...	1.425671e+09	9600013.0	148.507231
2	India	POLYGON ((97.32711 28.26158, 97.40256 27.88254...	1.428628e+09	3287259.0	434.595407
3	Kazakhstan	POLYGON ((87.35997 49.21498, 86.59878 48.54918...	1.960663e+07	2724902.0	7.195354
4	Saudi Arabia	POLYGON ((34.95604 29.35655, 36.06894 29.19749...	3.694702e+07	2149690.0	17.187141
5	Indonesia	MULTIPOLYGON (((141.00021 -2.60015, 141.01706 ...	2.775341e+08	1916862.0	144.785656

Simple map of Asia

Let's start by a creating a simple version of our chart:

create a figure and axe using the figure() and add_subplot() functions
create the cartogram with the cartogram() function from geoplot. We specify that we want the size and the color of each country to be mapped with the pop_norm_surface column of our dataset (aka density population)
create the background map with the popyplot() function from geoplot

And that's it!

fig = plt.figure(figsize=(12, 7))
ax = fig.add_subplot(111, projection=gcrs.PlateCarree())

gplt.cartogram(
   data, projection=gcrs.PlateCarree(),
   scale='pop_norm_surface', hue='pop_norm_surface', limits=(0,1),
   ax=ax
)
gplt.polyplot(data, ax=ax)
plt.show()

Custom colors

Now we can add a bit of customization:

load a color map using pypalettes
use the set_facecolor() function to change the background color of the graph
change the color of the background map
reduce the linewidth argument from 1 to 0.1

# colors
cmap = load_cmap("Antennarius_multiocellatus", cmap_type='continuous', reverse=True)
background_color = '#edf2f4'
text_color = '#14213d'
map_color = 'white'

fig = plt.figure(figsize=(12, 7))
ax = fig.add_subplot(111, projection=gcrs.PlateCarree())
fig.set_facecolor(background_color)
ax.set_facecolor(background_color)

gplt.cartogram(
   data, projection=gcrs.PlateCarree(), cmap=cmap,
   scale='pop_norm_surface', hue='pop_norm_surface', limits=(0,1),
   ax=ax
)
gplt.polyplot(data, facecolor=map_color, edgecolor='black', linewidth=0.1, ax=ax)
plt.show()

Title, subtitle and source

Now we need to add a bit of explanation about the chart:

we load custom fonts. Learn more about it this post
we use the fig_text() function from highlight_text to add the title, subtitle and source

# load the fonts
personal_path = '/Users/josephbarbier/Library/Fonts/' # change this to your own path
other_font = FontProperties(fname=personal_path + 'FiraSans-Light.ttf')
other_bold_font = FontProperties(fname=personal_path + 'FiraSans-Medium.ttf')

# colors
cmap = load_cmap("Antennarius_multiocellatus", cmap_type='continuous', reverse=True)
background_color = '#edf2f4'
text_color = '#14213d'
map_color = 'white'

# initiate figure and axes
fig = plt.figure(figsize=(12, 7))
ax = fig.add_subplot(111, projection=gcrs.PlateCarree())
fig.set_facecolor(background_color)
ax.set_facecolor(background_color)

# create the cartogram and background map
gplt.cartogram(
   data, projection=gcrs.PlateCarree(), cmap=cmap,
   scale='pop_norm_surface', hue='pop_norm_surface', limits=(0,1),
   ax=ax
)
gplt.polyplot(data, facecolor=map_color, edgecolor='black', linewidth=0.12, ax=ax)

fig_text( # title
   x=0.5, y=0.92, s="Population density (sq. km) in Asia",
   fontsize=25, ha='center', font=other_bold_font, color=text_color
)
fig_text( # subtitle
   x=0.5, y=0.86, s="Each country is the same size as it would be if it had the same population density as India.",
   fontsize=14, ha='center', font=other_font, color=text_color
)
fig_text( # credit and source
   x=0.7, y=0.13, s="<Data>: United Nations\n<Viz>: barbierjoseph.com",
   font=other_font, fontsize=7, color=text_color,
   highlight_textprops=[{'font': other_bold_font}, {'font': other_bold_font}]
)

plt.show()

Annotations of countries

data_projected = data.to_crs(epsg=4326): This line is converting the geospatial data to a common coordinate system (WGS84, used by GPS) for easier manipulation.
data_projected['centroid'] = data_projected.geometry.centroid: This line is calculating the centroid (geometric center) of each geometry in the data and storing it in a new column 'centroid'.
data['centroid'] = data_projected['centroid'].to_crs(data.crs): This line is converting the centroids back to the original coordinate system of the data.
The countries list contains the names of countries to be annotated.
The for loop iterates over each country in the countries list. For each country, it finds the centroid, adjusts its position if necessary, retrieves a value associated with the country, and then adds a text annotation at the adjusted centroid position on a map (not shown in the code).

# adjustement mapping for label positions
adjustments = {
   'Japan': (6, 0),
   'Philippines': (8, 0),
   'Sri Lanka': (0, -3.5),
   'Turkey': (-1, 1.2),
   'China': (0, -1),
   'Iran': (0, 2.4)
}

# load the fonts
personal_path = '/Users/josephbarbier/Library/Fonts/' # change this to your own path
other_font = FontProperties(fname=personal_path + 'FiraSans-Light.ttf')
other_bold_font = FontProperties(fname=personal_path + 'FiraSans-Medium.ttf')

# colors
cmap = load_cmap("Antennarius_multiocellatus", cmap_type='continuous', reverse=True)
background_color = '#edf2f4'
text_color = '#14213d'
map_color = 'white'

# create a figure object
fig = plt.figure(figsize=(12, 7))
ax = fig.add_subplot(111, projection=gcrs.PlateCarree())
fig.set_facecolor(background_color)
ax.set_facecolor(background_color)

# Generate the cartogram
gplt.cartogram(
   data, projection=gcrs.PlateCarree(), cmap=cmap,
   scale='pop_norm_surface', hue='pop_norm_surface', limits=(0,1),
   ax=ax
)
gplt.polyplot(data, facecolor=map_color, edgecolor='black', linewidth=0.1, ax=ax)

# get the centroids
import warnings ; warnings.filterwarnings("ignore") # mask warning about geometry attribute
data_projected = data.to_crs(epsg=4326)
data_projected['centroid'] = data_projected.geometry.centroid
data['centroid'] = data_projected['centroid'].to_crs(data.crs)
countries = ['China', 'India', 'Japan', 'Mongolia', 'Kazakhstan', 'Turkey', 'Philippines', 'Sri Lanka', 'Saudi Arabia', 'Iran']

# annotate each country
for country in countries:   
   centroid = data.loc[data['Country'] == country, 'centroid'].values[0]
   x, y = centroid.coords[0]
   x, y = (x + adjustments[country][0], y + adjustments[country][1]) if country in adjustments else (x, y)
   value = data.loc[data['Country'] == country, 'pop_norm_surface'].values[0]
   color = 'white' if country=='India' else text_color
   ax_text(
      x=x, y=y, s=f"<{country.upper()}>: {value:.0f}", fontsize=6, font=other_font, color=color,
      ha='center', va='center', ax=ax, highlight_textprops=[{'font': other_bold_font}]
   )

fig_text( # title
   x=0.5, y=0.92, s="Population density (sq. km) in Asia",
   fontsize=25, ha='center', font=other_bold_font, color=text_color
)
fig_text( # subtitle
   x=0.5, y=0.86, s="Each country is the same size as it would be if it had the same population density as India.",
   fontsize=14, ha='center', font=other_font, color=text_color
)
fig_text( # credit and source
   x=0.7, y=0.13, s="<Data>: United Nations\n<Viz>: barbierjoseph.com",
   font=other_font, fontsize=7, color=text_color,
   highlight_textprops=[{'font': other_bold_font}, {'font': other_bold_font}]
)

# save and show the plot
plt.savefig('../../static/graph/592-non-contiguous-cartogram-in-python.png', dpi=300)
plt.show()

Going further

Contact & Edit

👋 This document is a work by Yan Holtz. You can contribute on github, send me a feedback on twitter or subscribe to the newsletter to know when new examples are published! 🔥

This page is just a jupyter notebook, you can edit it here. Please help me making this website better 🙏!