[Python] Calculating the area of a radar chart over time by combining multiple arrays in Python

Suppose I have an array containing scores for n characters over t days. I have some more arrays of the same shape (t, n) containing different types of scores for the same n characters. For concreteness, let's say I have 5 scores, i.e. 5 arrays of shape (t, n). The individual scores (each normalized between 0 and 100) will change over time and my goal is to create a composite score using the area of the radar chart of the five scores for each character per day. (FYI: the area of the polygon is calculated by first assigning polar coordinates for each score (score_value, theta), converting these to (x, y) co-ordinates and then using a standard formula.) The final output containing the areas should also be a (t, n) array.

My initial idea was to combine the 5 scores into an array of shape (t, n, 5) achieved using np.dstack. Not sure if that's a good idea. Here's what I have so far:

import numpy as np

# Random data for demonstration
mat_1 = np.ones((20, 10))
mat_2 = np.ones((20, 10))
mat_3 = np.ones((20, 10))
mat_4 = np.ones((20, 10))
mat_5 = np.ones((20, 10))

matrix = [mat_1, mat_2, mat_3, mat_4]

my_list = np.dstack(matrix)

def radar_plot_area(data):
num_vars = len(data)

# Compute angle for each axis
angles = np.linspace(0, 2 * np.pi, num_vars, endpoint=False)

# Calculate area using shoelace formula
x = [r * np.cos(a) for r, a in zip(data, angles)]
y = [r * np.sin(a) for r, a in zip(data, angles)]
area = 0.5 * np.abs(np.sum(x * np.roll(y, -1)) - np.sum(x * np.roll(y, 1)))

return area

# Create output matrix
areas_matrix = np.zeros((20, 10))

for i in range(20):
for j in range(10):
areas_matrix[j] = radar_plot_area(my_list[j])

print(areas_matrix)


I can see it does output the correct values for the area at the end but I'm having trouble generalizing it to work with larger arrays. Ideally, I would prefer to work with arrays all the way throughout without having to explicitly refer to the indices too much, i.e. I want to have a function as follows:

def radar_plot_area(*scores: np.array): #the inputs are the 5 scores (which are arrays) I mentioned
...
return area #output is a (t, n) matrix containing the areas for each character per day


Is there a better way to achieve what I'm describing? Any suggestions would be much appreciated.

Continue reading...