Distance matrix of GeoDataFrames with scipy

gdf_distance_matrix.py

from scipy.spatial import distance_matrix
import pandas as pd

def gdf_distance_matrix(gdf1, gdf2, crs={'init':'epsg:3857'}):
    """
    Calculate the distance matrix between two GeoDataFrames
    Both GeoDataFrames must have the source crs set in order to be projected.

    Parameters
    ----------
    gdf1 : geopandas.GeoDataFrame
        GeoDataFrame #1
    gdf2 : geopandas.GeoDataFrame
        GeoDataFrame #2
    crs : str or dict
        Output projection parameters, passed to geopandas

    Returns
    -------
    pd.DataFrame
        Distance matrix dataframe; Distances can be looked up by .loc[]
    """

    ## Transform to mutual coordinate system to calculate distance
    dset1 = gdf1.to_crs(crs)
    dset2 = gdf2.to_crs(crs)
    
    ## List of coordinate pairs [x,y] for each dataset
    dset1_xy = dset1.apply(lambda b: [b.geometry.x, b.geometry.y], axis=1).tolist()
    dset2_xy = dset2.apply(lambda b: [b.geometry.x, b.geometry.y], axis=1).tolist()

    return pd.DataFrame(distance_matrix(dset1_xy, dset2_xy),
                        columns=dset2.index,
                        index=dset1.index)