Customer Personality Segmentation

Customer Personality Segmentation (MIT IDSS – Capstone Project)

As part of a machine learning capstone project from MIT-IDSS, a customer segmentation system was developed using a retail dataset containing demographic attributes, purchase behaviour, campaign responses, and spending habits.

The process began with data cleaning — handling duplicates, missing values, outliers, and inconsistencies — followed by univariate and bivariate analysis to explore patterns in customer behaviour.

Features were engineered from purchase history, campaign responses, and household data, and several clustering algorithms were implemented and compared:

K-Means
K-Medoids
Agglomerative Clustering
Gaussian Mixture Models

Silhouette scores, the elbow method, and cluster interpretability guided model selection. The final solution produced actionable customer profiles to support tailored marketing campaigns and retention strategies.

Tech Used: Python, Google Colab | Pandas, Scikit-learn and Seaborn (main libraries).

Notebook Preview

Customer segmentation notebook screenshot 1

Customer segmentation notebook screenshot 2

Customer segmentation notebook screenshot 3

Customer segmentation notebook screenshot 4

Due to MIT-IDSS copyright restrictions, the notebook cannot be shared publicly. I am happy to discuss further details upon request.