Project detail
Customer Segmentation with PyMC
Bayesian segmentation to group customers and prioritize sales actions with FastAPI and Flutter.
Highlights
- Bayesian customer clustering
- Weights, responsibilities, and cluster centers
- Visual app for exploring results
- Useful for commercial prioritization
No public repository or demo link is published for this project yet.
Architecture
FastAPI backend exposing probabilistic segmentation and Flutter frontend for reviewing clusters and customer distribution
Challenges
- Choosing truly informative behavioral variables
- Representing partial cluster membership
- Making the result understandable for business
Outcome
A foundation for treating customers as affinity groups rather than a single homogeneous mass
Overview
Problem
Not all customers respond the same, and flat segmentation falls short for commercial decisions.
Approach
I used PyMC to model segment membership and expose the result as a consumable API.
Outcome
The workflow enables more precise prioritization of campaigns and messages.