geoplot

geoplot is a high-level Python geospatial plotting library designed as an extension to cartopy and matplotlib. It aims to provide a seaborn-like experience for geospatial data.

Key features include:

• High-level plotting API: Provides easy access to standard cartographic map types for most common use cases.
• Native projection support: Simplifies the process of choosing and applying appropriate map projections for different types of geospatial depictions.
• Compatibility with matplotlib: Leverages the matplotlib ecosystem, making it easy to integrate with other Python visualization tools.

Install geoplot using the conda-forge channel to ensure all geospatial dependencies are correctly resolved.

conda install geoplot -c conda-forge

The recommended way to install geoplot is via Conda, which ensures compatibility across Linux, macOS, and Windows. Ensure you have Conda installed before running the command.

conda install geoplot -c conda-forge

You can install geoplot using pip. Note that this method may not work on Windows and might have compatibility issues on macOS and Linux.

pip install geoplot

If you have a standard pandas DataFrame containing latitude and longitude columns (e.g., from a CSV), you can convert it to a GeoDataFrame by using shapely.geometry.Point and the pandas.apply method.

import pandas as pd
import geopandas as gpd
from shapely.geometry import Point

# Load standard DataFrame
nyc_collisions_sample = pd.read_csv(gplt.datasets.get_path('nyc_collisions_sample'))

# Create geometry column from lat/long
collision_points = nyc_collisions_sample.apply(
    lambda srs: Point(float(srs['LONGITUDE']), float(srs['LATITUDE'])),
    axis='columns'
)

# Initialize GeoDataFrame
nyc_collisions_sample_geocoded = gpd.GeoDataFrame(nyc_collisions_sample, geometry=collision_points)

You can stack multiple plots (e.g., a polyplot of regions and a pointplot of cities) on top of each other by passing the axis object (ax) of the first plot to the subsequent plotting calls.

By default, geoplot sets the plot extent to the total_bounds of the last stacked plot. If you want to constrain the view to a specific area (like the contiguous US instead of the whole country), use the extent parameter with the total_bounds of your base geometry.

# Create the base projected map
ax = gplt.polyplot(
    contiguous_usa, 
    projection=gcrs.AlbersEqualArea()
)

# Stack a pointplot on the same axis, constraining the extent
gplt.pointplot(cities, ax=ax, extent=contiguous_usa.total_bounds)

By default, geoplot produces unprojected plots that treat coordinates as if they were on a flat Cartesian plane. To use a map projection, pass a geoplot.crs object to the projection parameter of a plotting function (e.g., polyplot, pointplot).

geoplot relies on cartopy for its projections. You can find a full list of available projections in the Cartopy documentation.

import geoplot.crs as gcrs
import geoplot as gplt

# Apply Albers Equal Area projection to a polyplot
gplt.polyplot(contiguous_usa, projection=gcrs.AlbersEqualArea())

To map data that lacks geometry (e.g., statistics by state), you must join that data against a GeoDataFrame that contains the corresponding shapes. This is typically done by setting a common index (like a state name) on both objects and using the .join() method.

import geoplot.crs as gcrs

# obesity is a standard DataFrame, contiguous_usa is a GeoDataFrame
result = contiguous_usa.set_index('state').join(obesity.set_index('State'))

# Plot the joined result
gplt.cartogram(result, scale='Percent', projection=gcrs.AlbersEqualArea())