Mechovate

AI-Powered Citizen Science & Data Quality Assurance Platform

Mechovate is a robust platform designed to streamline the lifecycle of citizen science data. It integrates machine learning for automated validation, GraphRAG for contextual intelligence, and interactive voice agents for seamless data reporting.

---

🚀 Key Features

• 🤖 Automated Validation: Multi-layer outlier detection using Scikit-learn (ML) and Great Expectations (GX) to filter invalid observations.
• 🎙️ Voice Agent Integration: Real-time voice interaction via LiveKit, allowing users to submit data or query metrics hands-free.
• 🧠 Contextual Intelligence: AI-driven chat powered by GraphRAG (LlamaIndex + Google Gemini) for deep insights and document-based information retrieval.
• 📊 Interactive Dashboard: Real-time map visualization of observations, validity indicators, and data trends.
• 🕷️ Data Scraper: Integrated Scrapy crawler (AQICN) for grounding citizen data against official environmental sources.

---

🛠️ Tech Stack

Frontend

• Framework: Next.js 16 (App Router)
• UI/UX: Tailwind CSS 4, Radix UI, Lucide React
• Visualization: React-Leaflet (Maps), Recharts (Analytics)
• Voice/Video: LiveKit Client SDK

Backend

• Core: FastAPI (Python)
• Database: SQLite / SQLAlchemy
• AI/ML: Scikit-Learn, Google Gemini (Generative AI)
• RAG: LlamaIndex, NetworkX
• Data Scraping: Scrapy

---

🏁 Getting Started

1. Prerequisites

• Node.js (v18+)
• Python (3.10+)
• LiveKit & Gemini API Keys

2. Backend Setup

cd backend
python -m venv .venv
# Activate venv:
# Windows: .venv\Scripts\activate
# Unix: source .venv/bin/activate
pip install -r requirements.txt
python main.py

3. Frontend Setup

cd frontend
npm install
npm run dev

4. Environment Configuration

Configure your .env files in both /frontend and /backend directories:

• LIVEKIT_API_KEY & LIVEKIT_API_SECRET
• GOOGLE_API_KEY (for Gemini)
• NEXT_PUBLIC_LIVEKIT_URL (for Frontend)

---

📂 Project Structure

• frontend/: Next.js application, components, and livekit integration.
• backend/: FastAPI server, ML validation logic, and AI services.
• backend/scraper/: Scrapy spiders for environmental data grounding.
• backend/knowledge_base/: Source materials for GraphRAG-based intelligence.

---

📄 License

Part of the Mechovate Citizen Science Initiative. Created for advanced data validation and interactive reporting.