Chatbot Groq API - Data Analysis Assistant

---

Open in Colab : https://colab.research.google.com/drive/1ldJH-h_yGU1itJUbYYSrERq7JQ-E1lL4?usp=sharing

---

Overview

This project is an interactive AI-powered chatbot that helps explore and analyze a Netflix dataset. It uses the Groq large language model API and vector search with FAISS to perform retrieval-augmented generation (RAG), enabling smart and context-aware answers about the data.

---

Features

• Data upload & chunking: Loads and splits large CSV datasets into manageable chunks for analysis.
• Semantic search: Uses embeddings from a sentence transformer model to retrieve relevant chunks of data based on user queries.
• Conversational analysis: Integrates with Groq’s LLM to generate detailed answers and insights in response to questions.
• Automated insights: Generates statistical summaries and trends from the dataset without manual prompting.

---

How It Works

• Dataset loading: Reads a CSV file containing Netflix titles, including metadata like type, title, director, cast, release year, rating, and description.
• Chunk creation: Splits the dataset into smaller JSON chunks for efficient retrieval.
• Embedding generation: Converts text chunks into numerical embeddings using a pretrained transformer.
• Vector indexing: Stores embeddings in a FAISS index to enable quick similarity search.
• Question answering: Retrieves the most relevant data chunks and sends them to the Groq LLM to generate natural language answers.
• Insight generation: Automatically summarizes dataset characteristics and extracts trends.

---

Tech Stack

• Groq API: For generating human-like responses.
• FAISS: For fast similarity search across large datasets.
• SentenceTransformers: To convert text into embeddings.
• Pandas & NumPy: For data manipulation and processing.

---

Project Structure

• Jupyter Notebook / Colab file containing:

  • Data preprocessing logic
  • Embedding and vector index creation
  • Functions to retrieve relevant chunks and generate answers
  • A conversational loop for interactive Q&A

• CSV dataset (Netflix titles) for analysis.

---

What You Can Do

• Ask questions about shows, movies, and dataset statistics.
• Discover trends, outliers, and summary statistics.
• Explore the dataset interactively, using natural language instead of manual filtering.

---

Requirements

• Python environment (e.g., Google Colab)
• Access to the Groq API (requires API key)

---

Usage Highlights

• Load a dataset and create embeddings.
• Ask questions about data (e.g., longest duration, earliest added show).
• Generate automated insights about distribution, release years, and other trends.

---