Create MultiPolygon from Dataframe column

Question

I have a dataframe with a column called geo_shape, for each row in this column data is in the format str with the following structure:

{"type":"MultiPolygon","coordinates":[[[[-99.1634043186724,19.3929362021363],[-99.1633935829153,19.3929814150842],[-99.1633779741792,19.3930471892216],[-99.1633676666966,19.3930905871479],[-99.1633539424123,19.3931484149288],[-99.1633536713463,19.393149556657],[-99.1633326516698,19.3932380913299],[-99.1633273408772,19.3932604765675],[-99.1633258846922,19.3932665988711],[-99.1632503982955,19.393250045508],[-99.1631580388255,19.3932297985256],[-99.1631022520036,19.3932175708424],[-99.1630096356198,19.3931972605697],[-99.1629243185542,19.3931785487074],[-99.1628253831388,19.3931568656532],[-99.1627721849032,19.3931451977826],[-99.1626860494573,19.3931263142213],[-99.1626865729325,19.3931240386763],[-99.162711364006,19.393124012126],[-99.1627524427473,19.3929332906781],[-99.1627565829766,19.3929140659265],[-99.1627814909565,19.3927984103957],[-99.1627980880081,19.3928020859532],[-99.1628764196311,19.3928194161424],[-99.1629163896067,19.3928282573258],[-99.1629931414886,19.3928452352628],[-99.1630458733238,19.3928569030601],[-99.1631201794492,19.3928733390959],[-99.1631648506849,19.3928832187645],[-99.1632414789013,19.3929001785344],[-99.1632705142618,19.3929065993979],[-99.1633527478204,19.3929247873036],[-99.1633697636313,19.3929285531163],[-99.1634043186724,19.3929362021363]]]]}

I want to convert each one of these rows to a MultiPolygon in a GeoDataframe. I have tried passing each str to a dict, then to a tuple and from there try to create the GeoDataframe, but my guess is that there is a quicker way of doing this.

Answer 1

Firstly, you have to create a shapely geometry by using shape function1

Secondly, you must use the geopandas library, specifying your dataframe and geometry column.

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

# First step
df['geo_shape'] = shape({"type":"MultiPolygon","coordinates":[[[[-99.1634043186724,19.3929362021363],[-99.1633935829153,19.3929814150842],[-99.1633779741792,19.3930471892216],[-99.1633676666966,19.3930905871479],[-99.1633539424123,19.3931484149288],[-99.1633536713463,19.393149556657],[-99.1633326516698,19.3932380913299],[-99.1633273408772,19.3932604765675],[-99.1633258846922,19.3932665988711],[-99.1632503982955,19.393250045508],[-99.1631580388255,19.3932297985256],[-99.1631022520036,19.3932175708424],[-99.1630096356198,19.3931972605697],[-99.1629243185542,19.3931785487074],[-99.1628253831388,19.3931568656532],[-99.1627721849032,19.3931451977826],[-99.1626860494573,19.3931263142213],[-99.1626865729325,19.3931240386763],[-99.162711364006,19.393124012126],[-99.1627524427473,19.3929332906781],[-99.1627565829766,19.3929140659265],[-99.1627814909565,19.3927984103957],[-99.1627980880081,19.3928020859532],[-99.1628764196311,19.3928194161424],[-99.1629163896067,19.3928282573258],[-99.1629931414886,19.3928452352628],[-99.1630458733238,19.3928569030601],[-99.1631201794492,19.3928733390959],[-99.1631648506849,19.3928832187645],[-99.1632414789013,19.3929001785344],[-99.1632705142618,19.3929065993979],[-99.1633527478204,19.3929247873036],[-99.1633697636313,19.3929285531163],[-99.1634043186724,19.3929362021363]]]]})
print(type(df)) # <class 'pandas.core.frame.DataFrame'>

# Second step
gdf = gpd.GeoDataFrame(df, geometry='geo_shape')
print(type(gdf)) # <class 'geopandas.geodataframe.GeoDataFrame'>

• https://shapely.readthedocs.io/en/stable/manual.html?highlight=import%20shape#shapely.geometry.shape

Answer 2

To create a shapely geometry, you can use shape function

from shapely.geometry import shape

shape({"type":"MultiPolygon","coordinates":[[[[-99.1634043186724,19.3929362021363],[-99.1633935829153,19.3929814150842],[-99.1633779741792,19.3930471892216],[-99.1633676666966,19.3930905871479],[-99.1633539424123,19.3931484149288],[-99.1633536713463,19.393149556657],[-99.1633326516698,19.3932380913299],[-99.1633273408772,19.3932604765675],[-99.1633258846922,19.3932665988711],[-99.1632503982955,19.393250045508],[-99.1631580388255,19.3932297985256],[-99.1631022520036,19.3932175708424],[-99.1630096356198,19.3931972605697],[-99.1629243185542,19.3931785487074],[-99.1628253831388,19.3931568656532],[-99.1627721849032,19.3931451977826],[-99.1626860494573,19.3931263142213],[-99.1626865729325,19.3931240386763],[-99.162711364006,19.393124012126],[-99.1627524427473,19.3929332906781],[-99.1627565829766,19.3929140659265],[-99.1627814909565,19.3927984103957],[-99.1627980880081,19.3928020859532],[-99.1628764196311,19.3928194161424],[-99.1629163896067,19.3928282573258],[-99.1629931414886,19.3928452352628],[-99.1630458733238,19.3928569030601],[-99.1631201794492,19.3928733390959],[-99.1631648506849,19.3928832187645],[-99.1632414789013,19.3929001785344],[-99.1632705142618,19.3929065993979],[-99.1633527478204,19.3929247873036],[-99.1633697636313,19.3929285531163],[-99.1634043186724,19.3929362021363]]]]})

For the whole DataFrame, you pass it to apply.

geometry = df['your_column'].apply(shape)