DataMind AI

DataMind AI is an open-source project dedicated to building an intelligent, dynamic dashboard for IoT data analysis. It leverages Generative AI to automatically summarize datasets, detect anomalies, predict machine failures, and generate beautiful, interactive visualizations.

🚀 Project Vision

We aim to solve the problem of visualizing and understanding large IoT datasets without manual configuration. By combining high-performance data processing (Pandas/Scikit-Learn) with the contextual intelligence of LLMs (Google Gemini), DataMind AI transforms raw CSV/Excel sheets into actionable insights.

🏗 Architecture

The project follows a modern decoupled architecture:

Backend (core/)

• Framework: FastAPI (Python)
• Data Processing: Pandas & Scikit-Learn
  • Hybrid Approach: We don't send massive datasets to the LLM. Instead, we extract metadata and statistical summaries locally.
  • Logic: The LLM acts as the "Brain" to define aggregation rules and anomaly detection parameters, while the Python backend acts as the "Muscle" to execute these rules on the full dataset.
• AI Integration: Google Gemini (via google-generativeai)
• Database: MongoDB (for storing file metadata and results)

Frontend (ui/)

• Framework: React + Vite
• Styling: TailwindCSS
• Visualization: Recharts (Dynamic rendering based on JSON config)
• State Management: Zustand

🌟 Key Features

1. 🧠 Smart Data Ingestion

• Universal Support: Drag-and-drop support for CSV, Excel (.xlsx), and JSON files.
• Large File Handling: Optimized to handle gigabyte-sized datasets by processing data locally using Pandas/Polars and only sending metadata to the cloud.

2. 🤖 AI-Driven Aggregation (The "Brain")

• Contextual Understanding: Gemini AI analyzes your column headers and data types to understand what your data represents (e.g., "This looks like IoT sensor data").
• Auto-Rule Generation: Instead of generic charts, the AI generates specific Pandas aggregation rules (e.g., "Group by DeviceID and calculate average Temperature") to reveal hidden patterns.

3. 🔍 Anomaly Detection & Predictive Maintenance

• Statistical Rigor: Uses Isolation Forests and Z-Score analysis to mathematically identify outliers.
• Context-Aware Alerts: The AI interprets why a data point is an anomaly (e.g., "Pressure dropped by 40% in 5 minutes") and suggests potential root causes.
• Failure Prediction: Aims to identify trendlines that historically precede machine failures.

4. 📊 Dynamic "Self-Constructing" Dashboard

• Zero-Config: You don't build charts; the AI builds them for you.
• Interactive UI: Fully responsive Recharts visualizations that support zooming, panning, and tooltips.

🐛 Issues & Feature Requests

Found a bug? Want to request a new feature?

1. Check the TODO.md to see if it's already planned.
2. Open a GitHub Issue with a clear title and description.
3. We welcome detailed bug reports!

🤝 Contribution & Status

This is a community-driven project. check TODO.md to see the progress and open tasks.

🛠️ Getting Started

Prerequisites

• Python 3.10+
• Node.js 18+
• MongoDB
• Google Gemini API Key

Backend Setup

1. Navigate to the core directory:

   cd core

2. Install dependencies:

   poetry install

3. Configure Environment Variables:

   • Copy the example environment file:

     cp .env.example .env

   • Open .env and add your keys:
     • MONGO_URI: Your MongoDB connection string.
     • GEMINI_API_KEY: Required. Get your free API key from Google AI Studio.

     ⚠️ Note: Never commit your .env file or API keys to GitHub!

4. Run the Server:

   poetry run uvicorn app.main:app --reload

Frontend Setup

cd ui
npm install
npm run dev

📜 License

MIT