I completed this project as part of the Machine Learning class at UC Berkeley alongside my peers Chloe Nguyen and Catherine Liao. I want to express my gratitude for their valuable contributions to the project!
This project focuses on conducting a comprehensive customer personality analysis for optimizing marketing strategies. Through extensive data preprocessing and clustering techniques such as KMeans and Agglomerative Clustering, we aim to identify four distinctive customer segments to empower the company with actionable insights to tailor products and targeted campaigns, thereby maximizing customer engagement and conversion rates.
Source: https://www.kaggle.com/datasets/imakash3011/customer-personality-analysis
- Mini Batch K-Means Clustering
- Agglomerative Clustering
- DBSCAN
- GMM
All models showed similar results.
Common personas among the models:
- Customer persona 1: Lowest income families (1-2 kids) with low spending habits across all products and all purchasing methods
- Customer persona 2: Moderate income families (1-2 kids) with moderate spending habits and high deals, stores and web purchases
- Customer persona 3: High income singles/ small families (0-1 kid) with high spending habits across all product categories- especially wine sales
![Screenshot 2024-04-26 at 10 57 30 AM](https://private-user-images.githubusercontent.com/130858240/326021083-841cccd2-7d26-4a97-87c5-44dbd7fe9f9a.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3MjU5NzE3ODMsIm5iZiI6MTcyNTk3MTQ4MywicGF0aCI6Ii8xMzA4NTgyNDAvMzI2MDIxMDgzLTg0MWNjY2QyLTdkMjYtNGE5Ny04N2M1LTQ0ZGJkN2ZlOWY5YS5wbmc_WC1BbXotQWxnb3JpdGhtPUFXUzQtSE1BQy1TSEEyNTYmWC1BbXotQ3JlZGVudGlhbD1BS0lBVkNPRFlMU0E1M1BRSzRaQSUyRjIwMjQwOTEwJTJGdXMtZWFzdC0xJTJGczMlMkZhd3M0X3JlcXVlc3QmWC1BbXotRGF0ZT0yMDI0MDkxMFQxMjMxMjNaJlgtQW16LUV4cGlyZXM9MzAwJlgtQW16LVNpZ25hdHVyZT01ZTgzNDk4MTNhYzI0MDE3ODRiODU5MGI2MWU3ZDZiZGRjNmE4ZDhiYmZkZTA4YTU2OTBhYWFkZjVkNDZiNDJlJlgtQW16LVNpZ25lZEhlYWRlcnM9aG9zdCZhY3Rvcl9pZD0wJmtleV9pZD0wJnJlcG9faWQ9MCJ9.KtytQoKp0t109BrkjTaKDXO4kNEMjpFKv6oNpbyxLQU)