GeoPandas

Warnings

• breaking GeoPandas 1.0 dropped support for Shapely < 2 and PyGEOS. It strictly requires Shapely >= 2. The `rtree` package for spatial indexing was also removed. If your project relied on older Shapely versions or PyGEOS, you must upgrade Shapely to version 2.0 or newer.
  Fix: Ensure `shapely>=2.0` is installed. Review and update code that explicitly used PyGEOS or `rtree` for spatial indexing.
• breaking The default I/O engine for reading and writing files (`gpd.read_file`, `gdf.to_file`) changed from Fiona to Pyogrio in GeoPandas 1.0. While `pyogrio` is now installed by default, ensure compatibility if you had specific configurations or relied on Fiona's unique behaviors.
  Fix: No direct code changes are usually needed, but be aware of potential subtle differences in behavior or performance. If explicit Fiona usage is desired, ensure it's installed and potentially specify `engine='fiona'` where applicable.
• breaking The `geopandas.datasets` module (e.g., `geopandas.datasets.get_path('naturalearth_lowres')`) has been removed in GeoPandas 1.0. Sample datasets are now available through the `geodatasets` library.
  Fix: Install `geodatasets` (`pip install geodatasets`) and update imports and calls from `geopandas.datasets.get_path` to `geodatasets.get_path`.
• deprecated Manually setting the Coordinate Reference System (CRS) via direct assignment (e.g., `gdf.crs = 'EPSG:4326'`) is deprecated. This method can lead to unexpected behavior or data corruption.
  Fix: Use the `GeoDataFrame.set_crs()` method (e.g., `gdf = gdf.set_crs('EPSG:4326')`) for setting the CRS. For re-projecting, use `GeoDataFrame.to_crs()`.
• gotcha Installing GeoPandas with `pip` can be challenging on some systems due to its dependencies on low-level C libraries like GEOS, GDAL, and PROJ. Missing or incorrectly compiled C libraries often lead to installation failures or import errors.
  Fix: The recommended installation method is using the `conda` package manager (e.g., `conda install -c conda-forge geopandas`), which provides pre-built binaries for these complex dependencies across different platforms. Using isolated virtual environments is also crucial.
• gotcha Spatial join operations (`gpd.sjoin()`) and spatial indexing (`sindex`) can fail with `RTreeError: Coordinates must not have minimums more than maximums` or similar errors if geometries are invalid (e.g., self-intersections, empty geometries) or if DataFrame indices are not clean and sequential.
  Fix: Validate geometries using `gdf.geometry.is_valid` and consider running `gdf.geometry.buffer(0)` or `gdf.geometry.make_valid()` to fix invalid geometries. Ensure DataFrame indices are reset with `df.reset_index(drop=True)` before performing spatial joins.
• breaking GeoPandas 1.1 requires Python 3.10 or greater, and updated minimum versions for core dependencies: pandas 2.0, numpy 1.24, and pyproj 3.5.
  Fix: Upgrade your Python environment to 3.10+ and ensure all listed core dependencies meet or exceed their minimum required versions.

Install

• pip install geopandas Basic Pip Install
• conda install -c conda-forge geopandas Recommended Conda Install (for C dependencies)

Imports

• GeoDataFrame

  from geopandas import GeoDataFrame

• GeoSeries

  from geopandas import GeoSeries

• read_file

  import geopandas as gpd
  gdf = gpd.read_file(...)

Quickstart

This quickstart demonstrates how to load a built-in geospatial dataset using `geodatasets` and `gpd.read_file()`, inspect its basic properties (like the first few rows and Coordinate Reference System), and create a simple static map using the `.plot()` method.

import geopandas as gpd
import geodatasets
import matplotlib.pyplot as plt

# Load a sample dataset (New York City boroughs)
path = geodatasets.get_path('ny.bb')
nybb = gpd.read_file(path)

print("GeoDataFrame Head:")
print(nybb.head())

print("\nCoordinate Reference System (CRS):")
print(nybb.crs)

# Plot the GeoDataFrame
fig, ax = plt.subplots(1, 1, figsize=(10, 10))
nybb.plot(ax=ax, color='lightgray', edgecolor='black')
ax.set_title('New York City Boroughs')
plt.show()