Data Analytics
KMeans
Clustering is the process of creating groups of objects based on the closeness (similarity) of the objects. K-means, DBSCAN, are some of the clustering algorithms.
Example below generates 2 clusters using kmeans algorithm. In the example income and spending are part of dictionary named "data". DataFrame is created using dictionary "data". KMeans is a python method, it is passed n_clusters (number of clusters) = 2, and random_state = 42. (random_state sets the random seed so that K-Means gives consistent and repeatable results across multiple runs. It ensures that the algorithm produces the same clustering result every time the program is run. Using a fixed value (e.g., random_state=42) makes the results reproducible, which is useful for experiments and debugging.)
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
if __name__ == "__main__":
data = {
'Income': [15, 16, 17, 18, 19, 20, 70, 75, 80, 85],
'Spending': [39, 40, 42, 43, 45, 47, 60, 65, 70, 75]
}
df = pd.DataFrame(data)
print(df)
#KMeans() creates a K-Means clustering model. n_clusters=2 tells the algorithm to divide the data into 2 clusters.
kmeans = KMeans(n_clusters=2, random_state=42)
#df[['Income', 'Spending']] selects the Income and Spending columns as input features.
#fit_predict() performs two tasks:
#fit(): Finds the cluster centers (centroids).
#predict(): Assigns each data point to the nearest cluster.
#The assigned cluster labels (0 or 1) are stored in a new column named Cluster in the DataFrame.
df['Cluster'] = kmeans.fit_predict(df[['Income', 'Spending']])
print("Cluster Centers:")
print(kmeans.cluster_centers_)
plt.scatter(df['Income'], df['Spending'], c=df['Cluster'])
plt.scatter(
kmeans.cluster_centers_[:, 0],
kmeans.cluster_centers_[:, 1],
s=200,
marker='X'
)
# : means select all rows.
# 0 means select the first column.
# : means all rows.
# 1 means the second column.
plt.xlabel('Income')
plt.ylabel('Spending Score')
plt.title('K-Means Clustering')
plt.show()