A data pipeline and analytics platform for understanding customer behavior — including segmentation, behavioral pattern analysis, and churn prediction.
Businesses collect customer data but rarely turn it into actionable intelligence — leaving churn, low engagement, and missed segments invisible until it's too late.
Built an end-to-end pipeline: data ingestion, feature engineering, clustering for segmentation, and a supervised model for churn prediction — surfaced through a reporting dashboard.
A working intelligence layer that gives a business real-time visibility into who their customers are, how they behave, and who's at risk of leaving.
My background in psychology shaped how I approached this: customer behavior isn't random — it follows patterns that, with the right data, are highly predictable. This project set out to make those patterns visible.
The platform combines unsupervised learning (K-means clustering) for customer segmentation with a supervised churn prediction model, all fed by a clean data pipeline that standardizes raw behavioral signals into usable features.
Add more context about your dataset, tools used, and any specific business outcomes here.
Data pipeline — Ingestion, cleaning, and feature engineering from raw transactional/behavioral data.
Segmentation — K-means clustering to identify distinct customer groups based on behavioral signals.
Churn prediction — Supervised classification model (add your model type here) trained on historical patterns.
Reporting layer — Dashboard for surfacing insights to non-technical stakeholders.