🚀 Tableau AI Copilot

Your intelligent AI assistant for Tableau Cloud - powered by Google Gemini and LangGraph agents

An AI-powered analytics platform that transforms how you interact with Tableau Cloud. Chat with your dashboards, discover visualizations through natural language, and get instant insights from your data using Google Gemini AI.

🔗 Quick Links

🌐 Live Demo: https://ai-insights-copilot.vercel.app/
💻 GitHub Repository: https://github.com/shalwin04/ai-insights-copilot
📡 Backend API: https://tableau-copilot-backend.onrender.com
🎥 Demo Video: https://youtu.be/n80pcvHKyI4
📚 Documentation: Setup Guide | Architecture

🎯 What Makes This Special?

Tableau AI Copilot bridges the gap between Tableau's powerful visualizations and natural language interaction. Instead of browsing through workbooks and views manually, simply ask questions and let AI find, display, and explain your dashboards.

Key Innovation:

🤖 AI-Powered Discovery: Semantic search across all your Tableau workbooks, views, and data sources
💬 Natural Language Chat: "Show me sales dashboard" → Instantly displays the right Tableau visualization
🎨 Embedded Visualizations: View and interact with Tableau dashboards directly in the chat interface
🔒 Secure Authentication: JWT-based authentication using Tableau Connected Apps

✨ Features

📊 Tableau Integration

Connected Apps JWT Authentication: Secure, server-side authentication with Tableau Cloud
Semantic Workbook Discovery: AI-powered search across workbooks, views, and data sources
Embedded Tableau Embedding API v3: Interactive dashboards rendered directly in the chat
Real-Time Indexing: Automatic workbook metadata indexing for instant discovery
Multi-Project Support: Search across all your Tableau projects and folders

🤖 Intelligent AI Agents (LangGraph)

Router Agent: Classifies user intent (Tableau discovery, data analysis, visualization)
Tableau Discovery Agent: Semantic search for Tableau content with OpenAI embeddings
Data Retriever Agent: Finds relevant datasets from Google Drive and other sources
Analyzer Agent: Plans and executes data analysis workflows
Summarizer Agent: Generates natural language insights and recommendations

📈 Data Analytics

Multi-Source Data Integration: Google Drive, CSV uploads, Excel files
Semantic Search: Vector-based search using OpenAI embeddings
Automated Workflows: Create pipelines from data ingestion to insights
Real-Time Insights: Live progress updates as agents analyze your data

💬 Natural Language Interface

Chat with Tableau dashboards using plain English
Context-aware conversations with memory
Real-time agent progress indicators
Session history persistence

🎨 Beautiful User Interface

Modern React UI with Tailwind CSS and shadcn/ui
Responsive design with dark mode support
Real-time WebSocket updates
Interactive data canvas for pinning insights

🏗️ Architecture

Frontend (React + TypeScript)
    ↓ REST API + WebSocket
Backend (Node.js + Express)
    ↓ LangGraph Multi-Agent Workflow
AI Agents (Router → Tableau Discovery → Analyzer → Summarizer)
    ↓ Google Gemini Pro + OpenAI Embeddings
Tableau Cloud (REST API v3.22 + Embedding API v3)
    ↓ Connected Apps JWT Auth
Vector Search (ChromaDB - Optional)
External Data Sources (Google Drive, Files)

Key Technologies:

AI/ML: LangChain.js, LangGraph, Google Gemini Pro, OpenAI Embeddings
Tableau: REST API v3.22, Embedding API v3, Connected Apps (Direct Trust)
Backend: Node.js, Express, TypeScript, Socket.IO
Frontend: React, TypeScript, Tailwind CSS, shadcn/ui, Framer Motion
Auth: OAuth 2.0 (Google), Tableau JWT (Connected Apps)
Vector DB: ChromaDB (optional for enhanced search)

🚀 Quick Start

Prerequisites

Node.js 18+ and npm
OpenAI API key (for embeddings)
Google Gemini API key
Tableau Cloud account with Connected Apps enabled
Google Cloud Console project (for Google Drive integration)

1. Clone the Repository

git clone <repository-url>
cd googlecloud-ai-accelerate

2. Configure Backend

cd Backend
npm install

# Create .env file with Tableau credentials
cat > .env << EOF
# Server
PORT=3000
SESSION_SECRET=your-secure-random-secret
FRONTEND_URL=

# AI Services
OPENAI_API_KEY=your-openai-api-key
GOOGLE_AI_API_KEY=your-gemini-api-key

# Tableau Cloud Connected Apps (Direct Trust)
TABLEAU_SERVER_URL=
TABLEAU_SITE_NAME=your-site-name
TABLEAU_USERNAME=your-tableau-username
TABLEAU_CLIENT_ID=your-connected-app-client-id
TABLEAU_SECRET_ID=your-connected-app-secret-id
TABLEAU_CLIENT_SECRET=your-connected-app-secret-value

# Optional: ChromaDB (for enhanced vector search)
CHROMA_ENABLED=false
CHROMA_URL=http://localhost:8000
EOF

Setting Up Tableau Connected Apps

Go to Tableau Cloud → Settings → Connected Apps
Create a new Connected App with Direct Trust
Enable scopes: tableau:content:read, tableau:workbooks:read, tableau:views:embed
Copy the Client ID, Secret ID, and Secret Value to your .env file

3. Configure Frontend

cd ../Frontend
npm install

# Create .env file
cat > .env << EOF
VITE_API_URL=http://localhost:3000
VITE_GOOGLE_CLIENT_ID=your-google-client-id
EOF

4. Start the Application

Terminal 1 (Backend):

cd Backend
npm run dev

Terminal 2 (Frontend):

cd Frontend
npm run dev

5. Open in Browser

http://localhost:5173

🎯 Usage

1. Connect to Tableau Cloud

Open the application
Navigate to the Tableau page (sidebar)
Click "Connect to Tableau Cloud"
Your workbooks will be automatically indexed

2. Browse Tableau Workbooks

View all your workbooks organized by project
Expand workbooks to see individual views
Click "Show Visualization" to embed dashboards
Interact with visualizations directly in the browser

3. Chat with Your Tableau Data

Ask natural language questions:

Tableau Discovery:

"Show me the sales dashboard"
"Find visualizations about revenue"
"What Tableau reports do I have for Q4?"
"Display the executive summary workbook"

Data Analysis:

"Analyze trends in my sales data"
"What are the key insights from Q4?"
"Compare revenue across regions"
"Summarize my customer data"

4. Watch the AI Agents Work

The interface shows real-time progress:

🤖 Router analyzes your intent
📊 Tableau Discovery searches for relevant dashboards
🔍 Retriever finds additional data if needed
📈 Analyzer generates insights
✍️ Summarizer creates natural language summaries

📖 Documentation

Setup Guide: Detailed installation and configuration
Architecture: System design and technical decisions
Tableau Setup: Tableau Connected Apps configuration

📊 API Endpoints

Tableau Endpoints

Endpoint	Method	Description
`/api/tableau/auth/connect`	POST	Connect with server-side credentials
`/api/tableau/auth/status`	GET	Check Tableau authentication status
`/api/tableau/auth/embedding-token`	GET	Get JWT token for embedding
`/api/tableau/workbooks`	GET	List all workbooks
`/api/tableau/views`	GET	List all views
`/api/tableau/views/:id/embed-url`	GET	Get embed URL for a view

Core Endpoints

Endpoint	Method	Description
`/api/auth/google`	GET	Initiate Google OAuth
`/api/drive/files`	GET	List Google Drive files
`/api/chat/message`	POST	Send chat message to AI agents
`/api/workflows`	GET	List automated workflows

🧪 Testing

Test Tableau Integration

cd Backend
npm run test:tableau

Test AI Workflow

npm run test:workflow

Manual Testing

# Connect to Tableau
curl -X POST http://localhost:3000/api/tableau/auth/connect

# Get workbooks
curl http://localhost:3000/api/tableau/workbooks \
  --cookie "connect.sid=your-session"

# Chat with Tableau
curl -X POST http://localhost:3000/api/chat/message \
  -H "Content-Type: application/json" \
  -d '{"message": "Show me sales dashboard"}' \
  --cookie "connect.sid=your-session"

🎨 Features Showcase

Semantic Search for Tableau

Uses OpenAI embeddings to create vector representations of:

Workbook names and descriptions
View names and content URLs
Data source metadata
Project names and tags

Embedded Visualizations

Tableau Embedding API v3 with:

JWT token authentication
Interactive filters and parameters
Full toolbar support (refresh, download, maximize)
Responsive sizing

Multi-Agent Workflow

LangGraph orchestrates:

Intent classification
Parallel data retrieval (Tableau + files)
Context-aware analysis
Natural language summarization

🐛 Troubleshooting

Tableau Connection Issues

# Verify credentials
curl -X POST http://localhost:3000/api/tableau/auth/connect

# Check environment variables
echo $TABLEAU_CLIENT_ID

Common Issues:

Invalid Connected App credentials → Verify Client ID and Secret in Tableau Cloud
Token expired → Server automatically refreshes tokens
Missing scopes → Ensure Connected App has read and embed permissions

Embedding Authentication Errors

If visualizations show "Sign in required":

Check that /api/tableau/auth/embedding-token returns a valid JWT
Verify Connected App has tableau:views:embed scope
Ensure TABLEAU_USERNAME matches your Tableau Cloud username

OpenAI/Gemini API Errors

Verify API keys in .env
Check API credits and rate limits
Gemini Pro requires billing enabled on Google Cloud

🗂️ Project Structure

tableau-ai-copilot/
├── Backend/
│   ├── src/
│   │   ├── agents/           # LangGraph AI agents
│   │   │   ├── tableauAgent.ts         # Tableau integration
│   │   │   ├── tableauDiscovery.ts     # Semantic search
│   │   │   └── workflow.ts             # Multi-agent orchestration
│   │   ├── services/
│   │   │   ├── tableau.ts              # Tableau REST API client
│   │   │   └── tableauJWT.ts           # JWT token generation
│   │   ├── routes/
│   │   │   └── tableau.ts              # Tableau API endpoints
│   │   └── config/
│   │       └── tableau.ts              # Tableau configuration
│   └── package.json
│
├── Frontend/
│   ├── src/
│   │   ├── components/
│   │   │   ├── features/
│   │   │   │   ├── TableauViz.tsx           # Embedding component
│   │   │   │   └── TableauWorkbookBrowser.tsx  # Workbook browser
│   │   │   └── pages/
│   │   │       └── TableauPage.tsx          # Main Tableau UI
│   │   ├── contexts/
│   │   │   └── TableauContext.tsx           # Tableau state management
│   │   └── lib/
│   │       └── api.ts                       # API client
│   └── package.json
│
└── README.md

🙏 Acknowledgments

Tableau for the powerful analytics platform and Embedding API
Google for Gemini Pro AI and Cloud infrastructure
LangChain.js for agent orchestration framework
OpenAI for embeddings API
shadcn/ui for beautiful UI components

Tech Stack: TypeScript · React · Node.js · Google Gemini · Tableau Cloud · LangGraph · OpenAI

Name		Name	Last commit message	Last commit date
Latest commit History 16 Commits
Backend		Backend
Frontend		Frontend
README.md		README.md

shalwin04/ai-insights-copilot

Folders and files

Latest commit

History

Repository files navigation