GeoPandas spatial join polygons within larger polygon/multipolygon

Question

I'm trying to spatially join smaller polygons that make up larger polygons/multipolygons using GeoPandas and the code below.

joined_gdf = gpd.sjoin(sua_2016_gdf, sa2_2021_gdf, how="left", predicate="intersects")

The sua_2016_gdf file is of my larger polygon boundaries and sa2_2021_gdf contains the smaller ones. The sa2 boundaries when combined will share a boundary with the sua ones.

I tried all the different predicate types (intersects, contains, within, touches, crosses, overlaps) but unfortunately the best I could manage was with intersects, which provided all the smaller polygons within, as well as all the smaller polygons touching the border outside, the larger polygon/multipolygon.

Is there a way I can strip out the smaller polygons that only touch the outside of the larger polygon's border so that only the polygons within remain?

Answer 1

You can buffer the small polygons by a small negative buffer distance so they fit inside the large polygon.

Or join by representative point.

import geopandas as gpd

smallpoly = gpd.read_file(r"C:\GIS\GIStest\smaller_polygon.shp")
largepoly = gpd.read_file(r"C:\GIS\GIStest\larger_polygon.shp")

sj1 = largepoly.sjoin(smallpoly, how="left", predicate="intersects")
#sj1.shape
#Out[19]: (45, 9)

#Buffer with a negative distance so the small polygons fit inside the large
smallpoly.geometry = smallpoly.buffer(-1)
sj2 = largepoly.sjoin(smallpoly, how="left", predicate="intersects")
# sj2.shape
# Out[24]: (26, 9)

#Or join by representative point.
smallpoly.geometry = smallpoly.geometry.representative_point()
sj3 = largepoly.sjoin(smallpoly, how="left", predicate="intersects")
#sj3.shape
#Out[30]: (26, 9)