Featured Projects

Challenge: Develop a machine learning model to predict NBA game outcomes with high accuracy using historical team performance data.

Technical Implementation: Built comprehensive data pipeline using Pandas for preprocessing, engineered 15+ statistical features, and implemented cross-validation to ensure model reliability.

Objective: Comprehensive performance comparison between DuckDB spatial extension, PostGIS and others for geospatial query processing on OpenStreetMap data.

Impact: Research findings influenced database selection for subsequent geospatial projects, providing actionable performance benchmarks for the development community.

Purpose: Create an engaging, responsive web application for interactive knowledge testing with real-time feedback and progress tracking.

Learning Focus: Mastered advanced DOM manipulation, state management without frameworks, and creating smooth user interactions through pure JavaScript.