Choropleths: Mapping data over space

Working with Geospatial Data in Python

Dani Arribas-Bel

Geographic Data Science Lab (University of Liverpool)

Choropleths

countries.plot(column='gdp_per_cap', legend=True)

Specifying a column:

locations.plot(column='variable')

Choropleth with classification scheme:

locations.plot(column='variable', scheme='quantiles', k=7, cmap='viridis')

Key choices:

locations.plot(column='variable', scheme='Quantiles', k=7, cmap='viridis')

Choropleths necessarily imply information loss (but that's OK)

Tension between:

Maintaining detail and granularity from original values (higher k)
Abstracting information so it is easier to process and interpret (lower k)

Rule of thumb: 3 to 12 classes or "bins"

locations.plot(column='variable', scheme='quantiles', k=7, cmap='viridis')

How do we allocate every value in our variable into one of the k groups?

Two (common) approaches for continuous variables:

locations.plot(column='variable', scheme='equal_interval', k=7, cmap='Purples')

locations.plot(column='variable', scheme='quantiles', k=7, cmap='Purples')

Categories, non-ordered

locations.plot(column='variable', 
               categorical=True, cmap='Purples')

Graduated, sequential

locations.plot(column='variable', 
               k=5, cmap='RdPu')

Graduated, divergent

locations.plot(column='variable', 
               k=5, cmap='RdYlGn')

IMPORTANT: Align with your purpose

Working with Geospatial Data in Python