Get the sum of an attribute within a set radius for each row using GeoPandas

Question

I have grid map that contains cells. Each cell is represented as a row in the dataframe, and each row contains the average population "pop" attribute.

What I need is that for each of these cells, calculate the sum of "pop" of every other cell including itself within a set radius (200m in this case), and assign that sum to a new attribute "coverage".

For example, the yellow dot will represent cell x. Cell x will be given a new attribute "coverage" which gets the sum of all "pop" including itself within the red circle. How do I then do this for all cells?

Answer 1

You can buffer your cells to create the coverage areas, then use a spatial join to link each coverage area with the cells it contains.

import geopandas
from shapely.geometry import Point

# create example data
cells = geopandas.GeoDataFrame({
    'index': [0,1,2,3],
    'pop': [1,2,3,4],
    'geometry': [
        Point([0,0]),
        Point([0,1]),
        Point([1,0]),
        Point([1,1])
    ]
}).set_index('index')

# create "coverage" buffers
coverage = cells[['geometry']].copy()
# take care here - your data should be in a projected CRS
# if you want to interpret this as a distance in meters
coverage['geometry'] = coverage.buffer(1.1) 

# join cells with coverage, then groupby and sum
cell_coverage_pop = cells.sjoin(coverage, predicate='within') \
    .groupby('index_right') \
    .sum()
# rename index and column name
cell_coverage_pop.index.rename('index', inplace=True)
cell_coverage_pop.rename(columns={'pop': 'coverage'}, inplace=True)
# cell_coverage_pop has "index" index and "coverage" 
# with population total

# join back to original cells if desired
cells = cells.join(cell_coverage_pop)
# cells now has "index" index, and original "pop", original 
# "geometry" and total "coverage"