Introduction to GeoPandas

Working with Geospatial Data in Python

Joris Van den Bossche

Open source software developer and teacher, GeoPandas maintainer

Spatial specific data formats

restaurants = pd.read_csv("datasets/paris_restaurants.csv")
restaurants.head()

                                             type         x          y
0                             Restaurant européen  259641.6  6251867.4
1                Restaurant traditionnel français  259572.3  6252030.2
2                Restaurant traditionnel français  259657.2  6252143.8
3  Restaurant indien, pakistanais et Moyen Orient  259684.4  6252203.6
4                Restaurant traditionnel français  259597.9  6252230.0

In the rest of the course:

spatial file formats (Shapefiles, GeoJSON, GeoPackage, ...)
GeoPandas: pandas dataframes with support for spatial data

Importing geospatial data with GeoPandas

import geopandas

countries = geopandas.read_file("countries.geojson")

countries.head()

          name      continent       gdp                            geometry
0  Afghanistan           Asia   64080.0  POLYGON ((61.21 35.65, 62.23 35...
1       Angola         Africa  189000.0  MULTIPOLYGON (((23.90 -11.72, 2...
2      Albania         Europe   33900.0  POLYGON ((21.02 40.84, 21.00 40...
4    Argentina  South America  879400.0  MULTIPOLYGON (((-66.96 -54.90, ...
5      Armenia           Asia   26300.0  POLYGON ((43.58 41.09, 44.97 41...

Quickly visualizing spatial data with GeoPandas

countries.plot()

The GeoDataFrame

countries.head()

          name      continent       gdp                            geometry
0  Afghanistan           Asia   64080.0  POLYGON ((61.21 35.65, 62.23 35...
1       Angola         Africa  189000.0  MULTIPOLYGON (((23.90 -11.72, 2...
2      Albania         Europe   33900.0  POLYGON ((21.02 40.84, 21.00 40...
...

type(countries)

geopandas.geodataframe.GeoDataFrame

The GeoDataFrame

countries.head()

          name      continent       gdp                            geometry
0  Afghanistan           Asia   64080.0  POLYGON ((61.21 35.65, 62.23 35...
1       Angola         Africa  189000.0  MULTIPOLYGON (((23.90 -11.72, 2...
2      Albania         Europe   33900.0  POLYGON ((21.02 40.84, 21.00 40...
...

A GeoDataFrame represents a tabular, geospatial vector dataset:

a 'geometry' column: that holds the geometry information
other columns: attributes describe each of the geometries

The 'geometry' attribute

countries.geometry

0      POLYGON ((61.21 35.65, 62.23 35...
1      MULTIPOLYGON (((23.90 -11.72, 2...
                      ...                
175    POLYGON ((23.22 -17.52, 22.56 -...
176    POLYGON ((29.43 -22.09, 28.79 -...
Name: geometry, Length: 176, dtype: object

type(countries.geometry)

geopandas.geoseries.GeoSeries

Spatial aware DataFrame

countries.geometry.area

0       63.593500
1      103.599439
2        3.185163
          ...    
174    112.718524
175     62.789498
176     32.280371
Length: 177, dtype: float64

Summary

A GeoDataFrame is like a pandas DataFrame:

all features of normal pandas DataFrames still work

but supercharged with spatial functionality:

plot() method
geometry attribute (GeoSeries)
spatial-specific attributes and methods (e.g. area)

Let's practice!

Working with Geospatial Data in Python