GPT-Powered Chatbot for Ride Sharing Data

• Tech Stack: Python, LangChain, OpenAI API, Streamlit, Pandas, Matplotlib
• Github URL: Codebase
• Link: FetiiBot

Natural Language Insights: Developed FetiiBot, a GPT-powered chatbot that answers user questions about Austin ride-sharing data, including trips, customers, and locations. Users can also request dynamic visualizations like age distributions or passenger counts.

Interactive Front-End & Context-Aware AI: Built with Streamlit for a smooth, interactive interface and LangChain + OpenAI GPT models for context-aware, multi-turn conversation handling, including session state management for seamless chat history.

Optimized Data Handling & Visualization: Merges and cleans multiple Excel datasets into a unified DataFrame. Visualization requests are dynamically generated using Matplotlib via a visualization agent, enabling on-demand charts while keeping API usage efficient.

Scalable & Extensible Design: Supports GPT model fallback to handle API token limits, structured for future real-time data pipelines, enhanced visualizations, personalized user queries, and a mobile-friendly, user-focused interface.