How can I extract the coordinates of an interior polygon using geopandas or shapely?

Question

I have a geojson file of county/unitary authorities (UTLAs) in the UK that I have read in using geopandas. I am trying to get a shapely polygon of each UTLA, but I am running into problems with UTLAs that contain other UTLAs (ie interior polygons), for example Derbyshire.

I am trying to extract the coordinates for the interior polygon:

import geopandas as gpd

utla_polygons = gpd.read_file('https://opendata.arcgis.com/datasets/244b257482da4778995cf11ff99e9997_0.geojson')

derbs = utla_polygons[utla_polygons['CTYUA21NM']=='Derbyshire']
derbs_int = derbs.explode().geometry.interiors
derbs_int

derbs_int is how I would expect it to be, based on the geopandas documentation :

130  0    [LINEARRING (-1.484521649999976 52.96638748100...
dtype: object

However, when I try to extract the coordinates from derbs_int, I get the following error:

derbs_int.coords

AttributeError: 'Series' object has no attribute 'coords'

How can I extract these coordinates so I can use them to create a shapely polygon?

Answer 1

One way to convert a multipoygon to a single polygon is by accessing the JSON and get the first element in the list.

import json
from shapely.geometry import shape
import requests

url = 'https://opendata.arcgis.com/datasets/244b257482da4778995cf11ff99e9997_0.geojson'
r = requests.get(url)
data = json.loads(r.text)
for f in data['features']:
    if f['properties']['CTYUA21NM'] == 'Derbyshire':
        geom = f['geometry']
        # convert MultiPolygon > Polygon
        geom = shape({
            'type': 'Polygon',
            'coordinates': geom['coordinates'][0]
        })
        print(type(geom))
        break

Output:

<class 'shapely.geometry.polygon.Polygon'>

Answer 2

Need to be systematic. Multi-polygons include polygons, polygons include interiors, interiors have co-ordinates

import geopandas as gpd
import requests

res = requests.get(
    "https://opendata.arcgis.com/datasets/244b257482da4778995cf11ff99e9997_0.geojson"
)
gdf = gpd.GeoDataFrame.from_features(res.json()).set_crs("epsg:4326")

gdfd = gdf.loc[gdf["CTYUA21NM"].str.contains("Derbyshire")].copy()

gdfd["geometry"].apply(
    lambda g: [g3.coords for g2 in g.geoms for g3 in g2.interiors]
).explode().explode()

output

130     (-1.484521649999976, 52.96638748100003)
130     (-1.484809956999925, 52.96630955400008)
130     (-1.484938351999972, 52.96627955300005)
130     (-1.485184568999955, 52.96623118800005)
130      (-1.48530681099993, 52.96621284400004)
                         ...                   
130     (-1.483622200999946, 52.96727346400007)
130     (-1.483387032999929, 52.96689850300004)
130    (-1.483327163999945, 52.966792199000054)
130    (-1.483532068999978, 52.966718497000045)
130     (-1.484521649999976, 52.96638748100003)
Name: geometry, Length: 1587, dtype: object