用Python 来做 Market Segmentation with Cluster Analysis

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
sns.set()
from sklearn.cluster import KMeans

data = pd.read_csv('3.12. Example.csv')
print(data)

代码紧跟上面

# Plot the data
plt.scatter(data['Satisfaction'], data['Loyalty'])
plt.xlabel('Satisfaction')
plt.ylabel('Loyalty')
plt.show()

# Select the features
x = data.copy()

# Clustering
kmeans = KMeans(2)
kmeans.fit(x)

# Clustering results
clusters = x.copy()
clusters['cluster_pred'] = kmeans.fit_predict(x)

plt.scatter(clusters['Satisfaction'], clusters['Loyalty'], c=clusters['cluster_pred'], cmap='rainbow')
plt.xlable('Satisfaction')
plt.ylable('Loyalty')
plt.show()

# Standaridze the variables
from sklearn import preprocessing

x_scaled = preprocessing.scale(x)
print(x_scaled) # x_scaled contains the standardized 'Satisfaction' and the same values for 'Loyalty'

# Take advantage of the Elbow method
wcss = []
for i in range(1,10):
  kmeans = KMeans(i)
  kmeans.fit(x_scaled)
  wcss.append(kmeans.inertia_)

print(wcss)

plt.plot(range(1,10), wcss)
plt.xlabel('Number of clusters')
plt.ylabel('WCSS')

# Explore clustering solutions and select the number of clusters
kmeans_new = KMeans(2)
kmeans_new.fit(x_scaled)
clusters_new = x.copy()
clusters_new['cluster_pred'] = kmeans_new.fit_predict(x_scaled)
print(clusters_new)

plt.scatter(clusters_new['Satisfaction'], clusters_new['Loyalty'], c=clusters_new['cluster_pred'], cmap='rainbow')
plt.xlabel('Satisfaction')
plt.ylabel('Loyalty')
plt.show()
"""
We often choose to plot using the original values for clearer interpretability. Note: the discrepancy we observe here depends on the range of the axes, too. 
"""

# Explore clustering solutions and select the number of clusters
kmeans_new = KMeans(4)
kmeans_new.fit(x_scaled)
clusters_new = x.copy()
clusters_new['cluster_pred'] = kmeans_new.fit_predict(x_scaled)
print(clusters_new)

plt.scatter(clusters_new['Satisfaction'], clusters_new['Loyalty'], c=clusters_new['cluster_pred'], cmap='rainbow')
plt.xlabel('Satisfaction')
plt.ylabel('Loyalty')
plt.show()

1. Types of analysis:
   [1] Exploratory
   — Get acquainted with the data
   — Search for patterns
   — Plan
   [2] Confirmatory
   [3] Explanatory

2. There are two types of clustering: Flat and Hierarchical

3. K means is a flat method in the sense that there is no hierarchy but rather we choose the number of clusters and the magic happens the other type is herarchical.

4. There are two types of hierarchical clustering agglomerative (bottom-up) and divisive (Top-Down).

5. With k-means we can simulate this divisive technique and that’s what we did with the elbow method.

6. Agglomerated and divisive clustering should reach similar results but agglomerated is much easier to solve mathematically.

7. Dendrogram: This solution has been produced on the same dataset based on Longitude and Latitude and we have standardized the variables. By the way, standardization did not make a difference in this case.

8. The bigger the distance between two links, the bigger the difference in terms of the features.

9. The pros of the Dendrogram:
   [1] Hierarchical clustering shows all the possible linkages between clusters.
   [2] We understand the data much, much better
   [3] No need to preset the number of clusters (like with k-means)
   [4] Many methods to perform hierarchical clustering (Ward method)

10. The cons of the Dendrogram:
    [1] It is also one of the reasons why hierarchical clustering is far from amazing is scalability.
    [2] It is extremely computationally expensive.
    [3] The more observations there are the slower it gets.
    [4] K means hardly has this issues.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

data = pd.read_csv('Country clusters standardized.csv', index_col='Country)

x_scaled = data.copy()
x_scaled = x_scaled.drop(['Language'], axis = 1)
print(x_scaled)

接着上面代码

sns.clustermap(x_scaled, cmap='mako')