A collection of web applications, data analysis, machine learning, and BI projects — each delivering real, measurable impact.
Website company profile B2B premium untuk penyedia LED Videotron, Sound System, VOD, dan Multimedia Entertainment korporat. Dual theme system (light/dark) dengan CSS variable, B2B copywriting persuasif, floating multi-CS WhatsApp FAB, dan komponen trust-building untuk decision maker enterprise.
A complete e-commerce platform built with Next.js 15, TypeScript, Drizzle ORM, and PostgreSQL. Features include product catalog, cart, checkout flow, order management, and an admin dashboard for managing inventory and orders in real time.
Modern, responsive company profile website built with Next.js and Tailwind CSS. Includes animated sections, full SEO optimization, contact forms, and a CMS-powered blog. Achieved 95+ Lighthouse performance score across all metrics.
Interactive web-based analytics dashboard that visualizes business KPIs in real time. Built with React and D3.js, connected to a Laravel REST API and PostgreSQL database. Used by operations teams to monitor daily performance.
Comprehensive Power BI dashboard for a retailer with 200+ branches. Analyzed sales patterns, inventory, and regional performance in real time. Resulted in a 23% improvement in stock management efficiency and reduced monthly reporting time from 3 days to hours.
Machine learning model to predict e-commerce customer CLV using ensemble methods. 87.3% accuracy with Random Forest + XGBoost. Helped the marketing team allocate acquisition budget 40% more efficiently and improved campaign ROI by 31%.
In-depth churn analysis for a telecom company with 500K+ customers using RFM Segmentation and predictive modeling. Identified 12 key churn factors and developed a data-driven retention strategy that reduced churn rate by 18% in 6 months.
Automated ETL pipeline integrating supply chain data from 12 different systems (ERP, WMS, TMS). Real-time monitoring dashboard with threshold alerts that reduced bottleneck identification time from 2 days to 2 hours.
Weekly sales prediction model using SARIMA and Prophet for a minimarket chain with 50+ outlets. MAPE below 8% with a 12-week prediction horizon. Reduced food waste by 19% and improved product availability by 14%.