Simple k means implementation to solve Coursera's quiz

k_means.py

import math
import statistics
import copy

centroids = [(2, 2), (-2, -2)]
data_points = [ \
(-1.88, 2.05), \
(-0.71, 0.42), \
(2.41, -0.67), \
(1.85, -3.80), \
(-3.69, -1.33)]

belongs_to = [0] * len(data_points)

# counts how many times a point has changed to answer
# Coursera's quiz question
changed = [0] * len(data_points)

def distance(data_point, centroid):
	return math.sqrt(sum([(data_point[i] - centroid[i]) ** 2 for i in range(len(data_point))]))

def find_closest_centroid_index_to_data_point(i):
	min_distance = float("+inf")
	min_index = 0
	for j in range(len(centroids)):
		dist = distance(data_points[i], centroids[j])
		if dist < min_distance:
			min_distance = dist
			min_index = j
	return min_index

def assign_data_point_to_nearest_centroid(i):
	closest_centroid_index = find_closest_centroid_index_to_data_point(i)
	if belongs_to[i] != closest_centroid_index:
		changed[i] += 1
	belongs_to[i] = closest_centroid_index

def assign_data_points_to_nearest_centroid():
	for i in range(len(data_points)):
		assign_data_point_to_nearest_centroid(i)

converged = False

def adjust_centers():
	before_update_centroids = copy.deepcopy(centroids)
	for i in range(len(centroids)):
		data_points_in_this_centroid = [data_points[j] for j in range(len(data_points)) if belongs_to[j] == i]
		x_mean = statistics.mean(map(lambda x: x[0], data_points_in_this_centroid))
		y_mean = statistics.mean(map(lambda x: x[1], data_points_in_this_centroid))
		centroids[i] = (x_mean, y_mean)
	global converged
	if centroids == before_update_centroids:
		converged = True

def k_means():
	while not converged:
		assign_data_points_to_nearest_centroid()
		adjust_centers()

def show_most_changed_data_point():
	max_index = 0
	for i in range(len(changed)):
		if changed[max_index] < changed[i]:
			max_index = i
	print("Data point {} changed the most; it changed {} times".format(max_index + 1, changed[max_index]))

if __name__ == "__main__":
	k_means()
	show_most_changed_data_point()

It uses Python 3