Scribd
Upload
44 views5 pages

Import Libraries

This document performs k-means clustering on location data (longitude and latitude) of customers in Hong Kong. It first loads and cleans the data, removing outliers. It then applies the elbow method to determine the optimal number of clusters is 4. K-means clustering with 4 clusters is applied and the centroids are identified. Distances from each point to centroids are calculated and each point is assigned to the closest cluster. The clustered points and centroids are plotted on a map with different colors.

Uploaded by

Mugabe Deogratious
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
44 views5 pages

Import Libraries

This document performs k-means clustering on location data (longitude and latitude) of customers in Hong Kong. It first loads and cleans the data, removing outliers. It then applies the elbow method to determine the optimal number of clusters is 4. K-means clustering with 4 clusters is applied and the centroids are identified. Distances from each point to centroids are calculated and each point is assigned to the closest cluster. The clustered points and centroids are plotted on a map with different colors.

Uploaded by

Mugabe Deogratious
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

Import Libraries

#-*- coding: utf-8 -*-#

# Pandas and numpy for data manipulation


import pandas as pd
import numpy as np
np.random.seed(8)

# Matplotlib and seaborn for plotting


%matplotlib inline
import matplotlib
import matplotlib.pyplot as plt
matplotlib.rcParams['font.size'] = 24
matplotlib.rcParams['figure.figsize'] = (16, 9)
plt.style.use('ggplot')

import seaborn as sns

from sklearn.cluster import KMeans

import warnings
warnings.filterwarnings('ignore')

Load Dataset
#set working directory
df = pd.read_csv("customer_location_df.txt", sep='|')
df.shape
(791776, 26)

Select Column
df = df[(df['GOO_COUNTRY_CODE']=='Hong Kong') & (df['PLC_STATUS']=='有效')]
df = df.loc[:,['GOO_LATITUDE','GOO_LONGITUDE']]

Check Missing Data


def check_missing_data(df):
# missing data
total = df.isnull().sum().sort_values(ascending=False)
percent = (df.isnull().sum() /
df.isnull().count()).sort_values(ascending=False)
missing_data = pd.concat([total, percent], axis=1, keys=['Total',
'Percent'])
print(missing_data.head(20))
check_missing_data(df)
Total Percent
GOO_LONGITUDE 0 0.0
GOO_LATITUDE 0 0.0
df = df.dropna()
df.describe()
GOO_LATITUDE GOO_LONGITUDE
count 281878.000000 281878.000000
mean 22.351708 114.150927
std 0.088876 0.338782
min 0.000000 0.000000
25% 22.305566 114.126470
50% 22.339085 114.171850
75% 22.388099 114.213265
max 28.647890 180.000000
df.head()
GOO_LATITUDE GOO_LONGITUDE
2 22.424867 114.233813
9 22.344887 114.194359
21 22.342677 114.166749
24 22.380442 114.172922
27 22.380882 114.199901
ax = sns.scatterplot(x="GOO_LONGITUDE", y="GOO_LATITUDE", data=df)

Outliner
df.loc[df['GOO_LATITUDE'] < 22]
GOO_LATITUDE GOO_LONGITUDE
403267 0.0 0.0
GOO_LATITUDE GOO_LONGITUDE
790967 0.0 0.0
df.loc[df['GOO_LONGITUDE'] > 115]
GOO_LATITUDE GOO_LONGITUDE
519918 28.64789 180.0

Update the GOO_LATITUDE, GOO_LONGITUDE for Outliner


#403267 22.4303384,114.0891702
#790967 22.3445576,114.1882836
#519918 22.3975032,113.9760157

df.GOO_LATITUDE.loc[[403267]] = 22.4303384
df.GOO_LONGITUDE.loc[[403267]] = 114.0891702
df.GOO_LATITUDE.loc[[790967]] = 22.3445576
df.GOO_LONGITUDE.loc[[790967]] = 114.1882836
df.GOO_LATITUDE.loc[[519918]] = 22.3975032
df.GOO_LONGITUDE.loc[[519918]] = 113.9760157
plt.scatter(df['GOO_LONGITUDE'], df['GOO_LATITUDE'], c='green', s=7)
<matplotlib.collections.PathCollection at 0x7f29bf7dd550>

The Elbow Method


wcss = []

for i in range(1, 11):


kmeans = KMeans(n_clusters = i, init = 'k-means++', max_iter = 300, n_init
= 10, random_state = 0)
kmeans.fit(df)
wcss.append (kmeans.inertia_)
plt.plot(range(1, 11), wcss)
plt.title('The Elbow Method')
plt.xlabel('Number of clusters')
plt.ylabel('WCSS')
plt.show()

Applying k-means to the dataset


Kmean = KMeans(n_clusters=4)
Kmean.fit(df)
KMeans(algorithm='auto', copy_x=True, init='k-means++', max_iter=300,
n_clusters=4, n_init=10, n_jobs=None, precompute_distances='auto',
random_state=None, tol=0.0001, verbose=0)
Kmean.cluster_centers_
array([[ 22.43797896, 114.18428292],
[ 22.30146589, 114.21442467],
[ 22.4148498 , 113.99272852],
[ 22.33590727, 114.14365898]])
c1 = (114.18428292, 22.43797896)
c2 = (114.21442467, 22.30146589)
c3 = (113.99272852, 22.4148498)
c4 = (114.14365898, 22.33590727)
def calculate_distance(centroid, X, Y):
distances = []

# Unpack the x and y coordinates of the centroid


c_x, c_y = centroid

# Iterate over the data points and calculate the distance using the
# given formula
for x, y in list(zip(X, Y)):
root_diff_x = (x - c_x) ** 2
root_diff_y = (y - c_y) ** 2
distance = np.sqrt(root_diff_x + root_diff_y)
distances.append(distance)

return distances
# Calculate the distance and assign them to the DataFrame accordingly
df['C1_Distance'] = calculate_distance(c1, df['GOO_LONGITUDE'],
df['GOO_LATITUDE'])
df['C2_Distance'] = calculate_distance(c2, df['GOO_LONGITUDE'],
df['GOO_LATITUDE'])
df['C3_Distance'] = calculate_distance(c3, df['GOO_LONGITUDE'],
df['GOO_LATITUDE'])
df['C4_Distance'] = calculate_distance(c4, df['GOO_LONGITUDE'],
df['GOO_LATITUDE'])
# Get the minimum distance centroids
df['Cluster'] = df[['C1_Distance', 'C2_Distance', 'C3_Distance',
'C4_Distance']].apply(np.argmin, axis =1)
df['Cluster'] = df['Cluster'].map({'C1_Distance': 'C1', 'C2_Distance': 'C2',
'C3_Distance': 'C3', 'C4_Distance': 'C4'})
plt.scatter(df['GOO_LONGITUDE'], df['GOO_LATITUDE'], c=df['Cluster'], s=5)
plt.scatter(c1[0], c1[1], marker='*', s=100, c='yellow')
plt.scatter(c2[0], c2[1], marker='*', s=100, c='red')
plt.scatter(c3[0], c3[1], marker='*', s=100, c='green')
plt.scatter(c4[0], c4[1], marker='*', s=100, c='blue')
<matplotlib.collections.PathCollection at 0x7f29b5cdb650>

You might also like