How parveshahameds/clustify works
Project type and purpose
• From the file tree (app.py, templates/, model files) and README summary, this is a monolithic Python web application (likely Flask) that serves a customer-segmentation tool. It loads pre-trained ML artifacts (CatBoost and pickle models), offers an HTML interface for users to upload data, runs inference, and returns segmented results.
File structure
• Top level
– app.py: entry point, defines HTTP routes and controllers
– requirements.txt: Python dependencies (Flask, scikit-learn, catboost, etc.)
– vercel.json: deployment configuration for Vercel
– app.log: application log file
• Model artifacts
– customer_segmentation_catboost.cbm
– customer_segmentation_model.pkl
• Presentation layer
– templates/index.html: Jinja2 template for the upload form and results
README for additional insights
• Although the README describes React/Node.js, the file layout shows a pure Python app with server-rendered HTML. So treat it as a Flask + Jinja2 web UI with embedded ML inference.
Detail
• Startup sequence: app.py loads models at launch.
• Error path: exception → app.log.