Geopandas density raster creation

Question

I have a geopandas dataframe that contains points at arbitrary lat/lon points. I would like to create a raster dataset so that each point in the regular grid contains the number of elements of the dataframe contained in a circle of a ccertain radius. The difficulty is that this number should contain only unique numbers respect to a certain column.

bb is my dataframe,radius is the distance from the center of the cell for whihc I want the numbers.

I am currently doing this using this code:

   mappa = numpy.full((nrows, ncols), 0) 
   for col in range(ncols):
        for row in range(nrows):
            lat=row*cellsize+bbox[1]
            lon=col*cellsize+bbox[0]   
            ff=bb.cx[lon-radius:lon+radius,lat-radius:lat+radius]  #  those give the rows falling in the limitss
            mappa[row,col]=len(ff['flkey'].unique())/period     # with this i count only unique points respect to a column flkey

The dataframe bb contains a huge number of points so that the method is extremely slow.

I would like to improve performing the check only in points in which there are sufficient numbers as most of the points in the map there are no points. So I would like to generate a preliminary density map and then analyse only the lat/lon points in which the density is significant.

Is there a point density funciotn similar to the ESRI function PointDensity_sa ?

Thanks

Could show what ff['flkey'] looms like, because it's unclear what you mean with "The difficulty is that this number should contain only unique numbers respect to a certain column." Kernel density approximation is the general term to create a density grid. en.wikipedia.org/wiki/Kernel_density_estimation . Scikit learn has an implementation scikit-learn.org/stable/modules/generated/… which avoids your loop by using trees. — Victor Savenije
– Victor Savenije, Commented Oct 20, 2024 at 20:50
The ff is a dataframe containing million of records, each one at lat/lon and with a parameter flkey that is a string. I want to create a raster in which at each cell, i want to have the number of records in a certain radius around the cell center. The problem is that I want unique records respect to flkey and not all the records in the circle. The normall density estimation cannot compute the unique records but count all the records around — Alessandro Annunziato
– Alessandro Annunziato, Commented Oct 23, 2024 at 7:36
Posted a solution. I suggest to reword the title to reflect the actual difficult coding part for you (raster of number of unique values) — Victor Savenije
– Victor Savenije, Commented Oct 23, 2024 at 18:18

Victor Savenije · Accepted Answer · 2024-10-23 18:16:23Z

1

The following uses sklearn.neighbors.KDTree to make fast queries from cell midpoints to data points. It yields the number of unique values present from a certain radius to the grid cell midpoints.

import numpy as np
from sklearn.neighbors import KDTree

n = 100_00
cell_size = 0.01
radius = 0.05


coords = np.random.random_sample((n, 2))
values = np.random.randint(0, n//10, n)

x_axis = np.linspace(0, 1, int(1 / cell_size))
y_axis = np.linspace(0, 1, int(1 / cell_size))
xv, yv = np.meshgrid(x_axis, y_axis)
tree = KDTree(coords)

results = tree.query_radius(list(zip(xv.ravel(), yv.ravel())), radius)
unique_results = [len(np.unique(values[indices])) for indices in results]
grid_values = np.array(unique_results).reshape((len(x_axis), len(x_axis)))

This yields an raster like this:

answered Oct 23, 2024 at 18:16

Victor Savenije

4072 silver badges12 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Geopandas density raster creation

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related