Project 5

Folio

Database Architecture[ Apr 2026 - May 2026 ]

Overview

Engineered an advanced bookstore information system using NoSQL MongoDB architecture with dynamic data tiering, automated shelf promotion, and Bayesian weighted ratings — outperforming relational joins by 28.2x in detailed read operations.

Key Features

◆Hot/Cold shelf data separation dividing documents into full-schema and archived minimalist schema, reducing document size by 75%
◆Node.js background worker using node-cron for autonomous popularity-based shelf promotion tracking user click frequencies
◆Pre-aggregated Summary Blobs serving top categories instantly without heavy database aggregations
◆Bayesian 3D weighted rating system fused with Regex sentiment analysis and time-decay factors to combat rating manipulation
◆Comprehensive benchmarking with Python scripts and k6 concurrency load testing proving 28.2x latency improvement over Neon SQL