AI Dashboard — Agentic Data Intelligence Platform

Architecture

Upload File → Ingestion → Profiling → Cleaning → Semantic Model → KPIs → Dashboard
                                                        ↓
                               User Question → SQL Agent → PostgreSQL → Chart

Quick Start

Prerequisites

• Python 3.11+
• Docker Desktop
• Anthropic API key (https://console.anthropic.com)

Install & Run

# 1. Clone / unzip the project
cd ai-dashboard

# 2. Run setup (creates .env, installs deps, starts DB)
chmod +x scripts/setup.sh
./scripts/setup.sh

# 3. Start everything with Docker
docker compose up --build

# 4. Open browser
open http://localhost:8000

Manual Run (without Docker for API)

source venv/bin/activate
docker compose up -d postgres redis          # DB only
celery -A pipeline.worker worker --loglevel=info &   # Background worker
uvicorn api.main:app --reload --port 8000    # API server

Project Structure

ai-dashboard/
├── agents/
│   ├── llm_client.py        # Anthropic API wrapper
│   ├── ingestion_agent.py   # File → PostgreSQL
│   ├── profiling_agent.py   # Schema + quality analysis
│   ├── semantic_agent.py    # Column → plain English mapping
│   └── kpi_sql_agent.py     # KPI builder + SQL generator + chart config
├── api/
│   └── main.py              # FastAPI routes
├── config/
│   └── settings.py          # Environment config
├── db/
│   ├── connection.py        # DB helpers
│   └── init.sql             # Schema creation
├── frontend/
│   └── index.html           # Dashboard UI + chat window
├── pipeline/
│   └── worker.py            # Celery async tasks
├── scripts/
│   └── setup.sh             # One-command setup
├── docker-compose.yml
├── Dockerfile
├── requirements.txt
└── .env.example

API Endpoints

Method	Path	Description
POST	/api/datasets/upload	Upload CSV/Excel/JSON/Parquet
GET	/api/datasets/{id}/status	Poll pipeline status
GET	/api/datasets/{id}/kpis	List generated KPIs
GET	/api/datasets/{id}/kpis/{kpi_id}/data	Execute KPI → chart config
POST	/api/chat	Natural language → SQL → chart
GET	/api/chat/{session_id}/history	Chat history
POST	/api/feedback	Flag incorrect result

Environment Variables

Variable	Description
ANTHROPIC_API_KEY	Your Anthropic API key
POSTGRES_*	Database connection details
REDIS_URL	Redis connection for Celery
LLM_SMART_MODEL	Model for SQL + semantic agents
LLM_FAST_MODEL	Model for KPI + chart agents