How to Apply Proportional Partitions to a Flow Cartogram¶

You can combine the flow cartogram and proportional cartogram modules to show two variables at once: the flow cartogram resizes each region by one variable, and proportional partitions subdivide each region to show a second variable.

A typical use case: resize states by total population, then split each state by party vote share.

In [1]:

import matplotlib.pyplot as plt

import carto_flow.data as examples
import carto_flow.flow_cartogram as flow
import carto_flow.proportional_cartogram as prop

import matplotlib.pyplot as plt import carto_flow.data as examples import carto_flow.flow_cartogram as flow import carto_flow.proportional_cartogram as prop

Step 1: Create a flow cartogram¶

Start by morphing the geometries. We'll use US Census data with race/ethnicity breakdowns as our secondary variable.

In [2]:

gdf = examples.load_us_census(population=True, race=True)
gdf["Minorities"] = gdf["Black or African American"] + gdf["Asian"] + gdf["Hispanic or Latino"]

cartogram = flow.morph_gdf(gdf, "Minorities", options=flow.MorphOptions(grid_size=512))
cartogram_gdf = cartogram.to_geodataframe()

gdf = examples.load_us_census(population=True, race=True) gdf["Minorities"] = gdf["Black or African American"] + gdf["Asian"] + gdf["Hispanic or Latino"] cartogram = flow.morph_gdf(gdf, "Minorities", options=flow.MorphOptions(grid_size=512)) cartogram_gdf = cartogram.to_geodataframe()

Morph geometries:  49%|█████████████████                  | 243/500 [00:09<00:10, 24.87it/s, max=7.0%, mean=0.4% - converged]

Step 2: Apply proportional partitions¶

partition_geometries() takes the morphed GeoDataFrame and splits each geometry according to the given columns. Use normalization='row' so each region fills completely (no remainder).

In [3]:

race_columns = ["Black or African American", "Hispanic or Latino", "Asian"]

partitioned = prop.partition_geometries(
    cartogram_gdf,
    columns=race_columns,
    method="split",
    direction="horizontal",
    strategy="treemap",
    normalization="row",
    treemap_reference="mean",
)

race_columns = ["Black or African American", "Hispanic or Latino", "Asian"] partitioned = prop.partition_geometries( cartogram_gdf, columns=race_columns, method="split", direction="horizontal", strategy="treemap", normalization="row", treemap_reference="mean", )

Step 3: Visualize¶

plot_partitions() colors each partition by category. The result shows both variables: region size encodes population, and the colored splits encode the race/ethnicity breakdown.

In [4]:

from carto_flow.proportional_cartogram.visualization import plot_partitions

fig, ax = plt.subplots(figsize=(12, 7))

palette = {"Black or African American": "indigo", "Asian": "darkorchid", "Hispanic or Latino": "plum"}

result = plot_partitions(partitioned, color_by="category", edgecolor="white", palette=palette, ax=ax)

result.legend.set_loc("lower left")
ax.set(title="Population cartogram with race/ethnicity breakdown")
plt.tight_layout();

from carto_flow.proportional_cartogram.visualization import plot_partitions fig, ax = plt.subplots(figsize=(12, 7)) palette = {"Black or African American": "indigo", "Asian": "darkorchid", "Hispanic or Latino": "plum"} result = plot_partitions(partitioned, color_by="category", edgecolor="white", palette=palette, ax=ax) result.legend.set_loc("lower left") ax.set(title="Population cartogram with race/ethnicity breakdown") plt.tight_layout();