Hi there - have been looking around for a good implementation of the K function and your one looks great! I'm actually using it for the 1D case, and I think I might have spotted a small bug in your code here:

Should that read: radius**2?

Hi,

Your function seems very useful. I'm not sure I understand the difference between radius and bounding radius.

I think the radius is the distance around each point where I want to measure K. Bounding radius seems to be the radius of the study area, used for spatial randomness assumptions? (Not having a perfect circular area, I'm using the maximum axis distance between two observations)

Here's a small example of how I'm applying your function and how my data is:

import ripleyk

# p1 is list of coordinate pairs
print(p1[2:7])
xs = np.array([i[0] for i in p1])
ys = np.array([i[1] for i in p1])

print(f"Xrange is: {max(xs) - min(xs)}")
print(f"Yrange is: {max(ys) - min(ys)}")

radius = [10,20]
bounding_radius = 150

k = ripleyk.calculate_ripley(radius, bounding_radius, d1=xs, d2=ys, CSR_Normalise=True)

k

Result:

[(56, 113), (43, 106), (41, 117), (67, 59)]
Xrange is: 90
Yrange is: 146
[2827.433388230814, 10040.27366432983]

So the output is the K value corresponding to each radius in the input. Do you have an implementation of p-value calculation (or rough estimation)? This very interesting paper minimizes Monte Carlo simulations for p-val estimation by having a model of point distribution that accounts for intensity and study area size.

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0080914

Cheers,
Ricardo

[-9.290137551123609e-06, -0.00010544767161566743, -0.00039142698870800463, -0.0007282757869681578, -0.0013863220751694216, -0.002301632731796177, -0.002895761209242842, -0.004294205083374969, -0.005929855937486295, -0.007915443959695345]