Completed

Marketing campaign analysis

Compare the segments before recommending the campaign.

Used customer demographics, spending, engagement, and response data to compare clustering methods and make segment-level marketing recommendations.

Inspect the artifacts View repository Engineering decisions

My role: I completed the analysis and recommendations as an MIT Professional Education capstone.
Result & status: Produced customer-group interpretations and recommendations for premium offers, loyalty efforts, and incentive outreach.

This was an MIT Professional Education capstone. No campaign deployment or measured lift is claimed.

Public evidence

Inspect the work.

Each link opens the source file or working product described below.

Academic analysis

Customer-segmentation notebook

Contains preparation, feature analysis, clustering comparisons, visualizations, and marketing recommendations. This is coursework, with no deployed campaign or measured lift.

The problem

Which customer patterns are distinct enough to support different outreach?

Prepare

Clean and scale demographic, behavioral, and spending features.

Compare

Evaluate several clustering methods instead of accepting one output.

Interpret

Connect group behavior to plausible outreach choices.

Approach

Compare methods before naming groups.

Cleaned and explored the dataset, scaled features, applied PCA, and compared hierarchical clustering, K-means, DBSCAN, K-medoids, and Gaussian mixtures.

Engineering

Turn clusters into behavioral profiles.

Model selection considered the interpretability of the resulting groups as well as their statistical separation. The analysis then tied each segment to spending and engagement behavior.

Validation

Check the business interpretation.

Compared clustering outputs and their business interpretation before recommending a segmentation approach.

Tools & methods

Python · PCA · Clustering · Customer segmentation