SQL & NoSQL Projects

Database-driven solutions focused on analytics, performance optimization, and scalable data architectures across SQL and NoSQL systems.

Customer Segmentation with RFM Analysis

Built a customer segmentation framework using Recency, Frequency, and Monetary scoring to identify high-value and at-risk customers.

PostgreSQL · SQL CTEs · Window Functions

Enabled targeted marketing strategies by identifying VIP and churn-risk segments. Improved retention by ~25% and increased campaign ROI.

Implemented CTEs, CASE logic, percentile ranking, and window functions. Optimized queries with indexes and materialized views.

⭐ View on GitHub

Sales Performance Dashboard

Designed optimized SQL queries powering a Power BI dashboard for regional sales, KPIs, and trend analysis.

SQL Server · Power BI · Analytics

Reduced reporting time by 40% and improved executive decision-making with near real-time insights.

Used joins, GROUP BY, window functions, and stored procedures. Automated refresh pipelines for large datasets.

⭐ View on GitHub

Smart Inventory Management System

Developed a real-time inventory system with automated alerts, predictive restocking, and audit trails.

PostgreSQL · Triggers · Automation

Reduced inventory costs by 30% and eliminated stockouts through predictive alerts and monitoring.

Normalized schema, enforced constraints, built triggers and scheduled jobs for real-time alerts.

⭐ View on GitHub

Clickstream Analytics Engine

Analyzed large-scale clickstream data to uncover user journeys, engagement patterns, and drop-off points.

MongoDB · Aggregation Pipeline · NoSQL

Improved conversion rates by 15% using data-driven UX optimizations.

Processed 50M+ events using MongoDB aggregation pipelines, optimized indexing, and streaming ingestion.

⭐ View on GitHub

Financial Data Reconciliation System

Built an automated SQL-based reconciliation pipeline for audit-ready financial datasets.

SQL Server · Data Quality · Compliance

Achieved 99.9% data accuracy and reduced manual processing time by 80%.

Used ROW_NUMBER for deduplication, validation rules, views for auditors, and automated ETL checks.

⭐ View on GitHub