Alejandro RubioWorkcell

Projects

E-commerce Customer Segmentation (RFM)

RFM features with K-Means, K-Prototypes for mixed variables and DBSCAN for location, on 852,584 transactions.

Type

Data

Context

MSc in Artificial Intelligence, UNIR, coursework

Period

2026

Stack

Python · scikit-learn · kmodes · pandas

Area

[AI & machine learning]

What was built

    I built recency, frequency and monetary features per customer, picked two clusters by silhouette score and profiled them. K-Prototypes adds gender and device. DBSCAN on home coordinates is compared with the RFM segments.

    Facts

    ItemValue
    Transactions852,584
    K-Means segments23,610 and 26,632 customers

    No public repository for this project.