Data Architecture & BI Agent with MCP, n8n, and Superset
Next-gen analytics platform integrating LLMs via the Model Context Protocol (MCP) for natural language querying, with n8n workflow orchestration and Apache Superset visualization.
Technologies & Infrastructure Used
📐 System Architecture & Data Pipeline
Technical Diagram
Click on the diagram to expand in full resolution.
### 🎯 Business Problem & Architectural Goal
Modern decision-makers often face friction when querying enterprise databases, relying on manual SQL requests or rigid static dashboards. The goal of this project was to architect an autonomous Business Intelligence engine that allows natural language data interaction while maintaining enterprise data security and pipeline reliability.
---
### ⚙️ Pipeline & System Architecture
The entire stack is containerized using **Docker** and orchestrated in a **Linux Fedora** production-ready environment:
* **Workflow Orchestration (n8n):** Serves as the central event-driven engine, handling incoming webhooks, data transformations, and state management.
* **AI & LLM Integration (Model Context Protocol - MCP):** Implements the open MCP standard to safely expose database schemas and structured context to LLMs, enabling accurate SQL generation and natural language query execution.
* **Analytics & Serving Layer (Apache Superset & PostgreSQL):** Stores structured metrics and exposes interactive analytical views for end-user exploratory analysis.
---
### 📈 Key Engineering Outcomes
* **Zero-Latency Insights:** Reduced query-to-insight time by allowing non-technical stakeholders to ask natural language questions directly against structured schemas.
* **Reproducible Infrastructure:** Multi-container Docker deployment ensuring full isolation, environment consistency, and easy cloud migration.
* **Modern Stack Standards:** Combines deterministic ETL/ELT pipeline principles with modern Agentic AI workflows.