DashboardLLM

It generates charts based on the data using Vega-Lite JSON specifications, ensuring safe and efficient visualization without any malicious code execution.

Key Features

• CSV Processing: Reads and processes data from uploaded CSV files to generate insights.
• Chart Generation: Creates interactive charts using Vega-Lite JSON specifications, ensuring secure and reliable visualization.
• Retrieval-Augmented Generation (RAG): Employs a vector database (ChromaDB) to find the most relevant data chunks to answer user queries.
• Local LLMs: Powered by local language models via Ollama, ensuring privacy and control over the models used.

Tech Stack

• Backend & Orchestration: Python, LangChain
• Frontend: Streamlit
• LLM Serving: Ollama
• Models:
  • Generation: Qwen (or any other powerful chat model)
  • Embeddings: granite-embedding:latest
• Vector Database: ChromaDB
• CSV Processing: Pandas

How It Works

1. CSV Upload: Users upload a CSV file containing structured data.
2. Context Retrieval: Relevant data chunks are retrieved using ChromaDB.
3. Chart Generation: The system generates Vega-Lite JSON specifications based on the data and renders interactive charts.
4. Interactive Chat Interface: Users interact with the chatbot via a Streamlit-based UI.

Getting Started

1. Clone the repository:

   git clone https://github.com/Aeets6/dashboard-LLM.git
   cd dashboard-LLM

2. Install dependencies:

   pip install -r requirements.txt

3. Run the application:

   streamlit run main.py

4. Upload a CSV file and start interacting with the chatbot.

Security

DashboardLLM ensures safe operation by:

• Using Vega-Lite JSON specifications for chart generation, eliminating the risk of malicious code execution.
• Employing local language models for privacy and control.