Retail Purchase Intelligence System: Implementation, Experimental Evaluation, And Performance Analysis

Abstract

This paper presents the implementation and experimental evaluation of the Retail Purchase Intelligence System (RPIS), a web-based platform engineered to automate multi-source price comparison across major e-commerce websites. Building upon the architectural framework established in the first-semester paper, this continuation focuses on the actual construction, deployment, and rigorous performance testing of the system. The RPIS employs Python-based web scraping technologies — specifically BeautifulSoup, Requests, and Selenium — to extract real-time product pricing data from multiple online retail platforms. The extracted data undergoes normalization and storage in a structured MySQL relational database before being presented through a responsive web interface developed using HTML5, CSS3, and JavaScript. Experimental evaluations conducted across five major e-commerce portals demonstrated a price extraction accuracy of 94.7%, an average system response time of 3.2 seconds, and a data normalization success rate of 96.1%. The system significantly reduces consumer effort in price comparison while enabling intelligent, data-driven purchase decisions. Results validate the feasibility and effectiveness of automated retail intelligence in the modern digital commerce landscape.