Pandas Rolling with evolutive windows

Question

I would like to use a pandas way to iterate over a geodataframe to compute an equivalent to a cumsum but with union of polygon. I have a series of polygons, the first keeps the same geometry, the second is the union between the first and the second, the third the union between the 3 first ect.

I was thinking about using rolling tools but the windows seems to be constant. Do you have any idea how to deal with it using pandas ?

Thanks.

Reproductible example :

import shapely
import geopandas as gpd

p0 = shapely.geometry.Point([0, 0]).buffer(1)
p1 = shapely.geometry.Point([1, 1]).buffer(1)
p2 = shapely.geometry.Point([1, 0]).buffer(1)

gdf = gpd.GeoDataFrame({'name' : ['p0', 'p1', 'p2'], 'geometry' : [p0, p1, p2]}, crs = 'epsg:4326')

# what i would like

gdf['evolving_area'] = [shapely.unary_union([p0]).area, 
                        shapely.unary_union([p0, p1]).area, 
                        shapely.unary_union([p0, p1, p2]).area]
gdf['evolving_area']

Answer 1

One option would be to perform an expanding window calculation :

gdf["evolving_area"] = list(
    map(lambda p: p.unary_union.area, gdf["geometry"].expanding())
)

Output :

  name                                           geometry  evolving_area
0   p0  POLYGON ((1.00000 0.00000, 0.99518 -0.09802, 0...       3.136548
1   p1  POLYGON ((2.00000 1.00000, 1.99518 0.90198, 1....       5.704823
2   p2  POLYGON ((2.00000 0.00000, 1.99518 -0.09802, 1...       6.832249