Customer Segmentation Analysis

Overview | Key Features | Cluster Descriptions | How to Install | Credits

Overview

This script performs customer segmentation analysis using K-means clustering, an unsupervised machine learning (ML) technique, on a marketing dataset. The goal is to group customers based on their purchasing behavior and other features, providing valuable insights for targeted marketing strategies.

Key Features

• Data preprocessing to handle missing values and duplicates.
• Visualization of histograms to understand the distribution of features.
• Determination of the optimal number of clusters using the elbow method.
• K-means clustering to group customers into distinct segments.
• Saving cluster centers and summary to a file for further analysis.
• Visualizing clusters in a scatter plot using principal component analysis (PCA).

Cluster Descriptions

Cluster 0 (Careful Spenders):

Description: These customers are cautious with their spending and prefer to pay off their balances in full.

Key Characteristics:

• Low balance ($1500) and cash advance ($800).
• Low purchase frequency (0.16) and moderate credit limit (~$3300).
• Moderate payments made and minimum payments (~$975 and ~$790, respectively).
• Low percentage of full payment (2.18%).

Cluster 1 (Low Activity):

Description: Customers in this cluster have relatively low activity on their credit cards.

Key Characteristics:

• Low balance ($110) and cash advance ($335).
• Low purchase frequency (0.27) and credit limit (~$3740).
• Moderate payments made and low minimum payments (~$1094 and ~$268, respectively).
• Moderate percentage of full payment (23.09%).

Cluster 2 (Active Users):

Description: These customers are active users of their credit cards, making frequent purchases and paying off their balances.

Key Characteristics:

• Moderate balance ($1850) and cash advance ($336).
• High purchase frequency (0.93) and credit limit (~$7054).
• High payments made and moderate minimum payments (~$2884 and ~$807, respectively).
• High percentage of full payment (28.52%).

Cluster 3 (Heavy Cash Users):

Description: Customers in this cluster heavily rely on cash advances and have moderate purchase activity.

Key Characteristics:

• High balance ($4880) and cash advance ($5052).
• Moderate purchase frequency (0.31) and credit limit (~$8001).
• High payments made and high minimum payments (~$3961 and ~$2057, respectively).
• Low percentage of full payment (3.90%).

Cluster 4 (High Spenders):

Description: These customers are high spenders with substantial credit limits and frequent purchases.

Key Characteristics:

• High balance ($4692) and cash advance ($1011).
• Very high purchase frequency (0.92) and credit limit (~$12604).
• Very high payments made and high minimum payments (~$16307 and ~$3625, respectively).
• High percentage of full payment (38.96%).

Cluster 5 (Moderate Users):

Description: Customers in this cluster have moderate spending habits and tenure.

Key Characteristics:

• Moderate balance ($748) and cash advance ($212).
• High purchase frequency (0.88) and credit limit (~$3302).
• Moderate payments made and minimum payments (~$1009 and ~$735, respectively).
• Moderate percentage of full payment (25.15%).

How to Install

1. Install the required Python packages:

    pip install pandas numpy seaborn matplotlib scikit-learn jupyterthemes

2. Clone the repository:

    git clone https://github.com/nmcintyre5/customerSegmentation.git

3. Navigate to the project directory:

    cd customerSegmentation

4. Run the main script customer_segmentation.py:

    python customerSegmentation.py

5. View the generated plots and output files for insights into customer segmentation.

Credits

This script is based off of the Coursera course "Unsupervised Machine Learning for Customer Market Segmentation" by Ryan Ahmed, Adjunct Professor & AI Enthusiast.

Additional Tags

#CustomerSegmentation #DataAnalysis #Clustering #MarketingAnalytics #KMeans #PCA #DataVisualization