Back to all projects

Paris-Dauphine analytics and data visualization project

Customer Segmentation Dashboard

Clustering, profiling, and dashboarding for marketing decisions

This project combines customer analytics and product thinking. The modeling work identifies meaningful segments, while the dashboard layer makes the output usable by a non-technical team for campaign targeting and profiling.

Customer Segmentation Dashboard visual

2,240

Clients

35

Engineered vars

4

Clusters

Problem

Marketing teams need segments that are both statistically meaningful and easy to explore without reopening notebooks or raw data files.

Approach

Prepared the data, engineered customer behavior features, compared K-Means, hierarchical clustering, and GMM, selected four segments, and built a Shiny application with filters, KPIs, and radar profiles.

Results

Recovered four usable customer groups over 2,240 clients and turned the analysis into a live dashboard instead of a static report.

What is in the repository

Handled missing values and clear outliers before clustering.
Compared several unsupervised approaches instead of relying on a single algorithm.
Created radar-style segment profiles to support stakeholder discussion.
Delivered an interactive interface that moves the project closer to real usage.

Role and scope

Unsupervised modeling, feature engineering, visualization, and dashboard delivery

Project context

Paris-Dauphine analytics and data visualization project

Main stack

RShinytidyverseggplot2FactoMineRplotly