Real-Time Sentiment Analysis Pipeline for Yelp Customer Feedback Using ChatGPT's LLM and Distributed Systems

This work proposes a real-time sentiment analysis pipeline on customer feedback using Yelp and addresses the high-volume dynamic user-generated contents processing problem. The proposal integrates state-of-the-art machine learning models and distributed systems, making use of Large Language Models within a Docker container called ChatGPT for the real-time sentiments analysis. The system ingests live Yelp data streams through TCP/IP, processes them in a scalable Apache Spark cluster, and ensures that data flows seamlessly through Apache Kafka. Enriched feedback is indexed into Elasticsearch, providing powerful analytics and visualizations through tools like Kibana, Tableau, and Power BI. The low-latency design of the pipeline gives enterprises actionable insights for real-time decision-making, enabling enhanced customer sentiment analysis and business intelligence. Scalable and efficient, this architecture fits all large volume Enterprise Feedback Management (EFM) needs when looking to achieve insights fast from customer sentiment.