How globalmediaa/apriori-basket-analysis works
Project type and purpose
• This is a small monolithic Python “micro-service” that performs market-basket (sepet) analysis and serves results over HTTP.
• It uses Flask for the web layer, Pandas for data manipulation, and mlxtend for association-rule mining on a CSV data source.
File structure
• Top level:
• basket.py – main Flask application, ties together HTTP routes, data loading, and the analytics engine
• store_data.csv – static source of transaction data for analysis
• README.md – installation and run instructions
• No separate “frontend” directory—presentation logic is embedded in Flask routes/templates (if any). No external database; the CSV is the only data store.
README insights
• Dependencies: pandas, mlxtend, flask
• Run steps: start server on port 5000 and request via URL with a product_id query param
• Implied flow: user calls an endpoint, server reads CSV, computes basket rules, returns result
Main components and interactions
a. HTTP Client (browser or API consumer)
b. Flask Web Server (basket.py)
– Request Handler / Router
– Data Loader module (reads store_data.csv via Pandas)
– Analytics Engine (mlxtend rule-mining + Pandas transformation)
– Response Formatter (JSON or HTML)
c. Data Store (store_data.csv)
Relationships:
– Client → Flask Request Handler (HTTP GET /?id=product_id)
– Request Handler → Data Loader (reads CSV)
– Data Loader → Analytics Engine (feeds DataFrame)
– Analytics Engine → Response Formatter (associates rules + filters by product_id)
– Response Formatter → Client (HTTP response)
Architectural patterns and design principles
• Monolithic but logically layered: presentation (Flask routes), business logic (analytics), data access (CSV reader)
• Single process, single codebase, no external service dependencies
• Could evolve to a microservices setup by splitting data-access, analytics, and API layers later
Putting it all together
1. Request Handler (Flask route)
2. Data Loader (Pandas reading CSV)
3. Analytics Engine (mlxtend rule mining)
4. Response Formatter