Explore

N8N: Automated EV Utilisation Alerting System

Goal

To build a fully autonomous, real-time anomaly detection pipeline that uses Generative AI to analyse critical infrastructure data.

Technology Stack

Item

Tool

Detail

Data Engineering

PostgreSQL, Python (Pandas, SQLAlchemy)

Building a scalable, production-ready data backbone.

AI / NLP

OpenAI (GPT 4.1 mini)

Advanced Generative AI integration for data analysis and message synthesis.

Automation

n8n

End-to-end process scheduling and execution flow control.

Messaging

Gmail API (OAuth 2.0)

Secure, external API integration, overcoming complex authentication hurdles.

There are no rows in this table

⁠

Pipeline & Architecture (How It Works)

⁠

⁠

Database creation

Python script to create SQL database (PostgreSQL)

⁠

Data Ingestion (PostgreSQL)

Python scripts simulate real-time EV station utilisation data, ingesting it into a PostgreSQL database where it’s tracked in the utilisation and alerts tables.

Generate data

⁠

Check anomaly

⁠

Connect database with Beekeeper tool

⁠

Test the SQL query

⁠

N8N Set Up

⁠

Orchestration Trigger (n8n)

The system is kicked off on a schedule (simulating a Cron Job) to check for new data.

Anomaly Detection & Filtering

The PostgreSQL node queries the database for all new utilisation events exceeding the 70% threshold. An IF Node validates that at least one alert exists.

Intelligent Message Generation (OpenAI)

The single JSON object is passed to the OpenAI Node.

A specific System/User Prompt is used to instruct the LLM to analyse the raw data, prioritise CRITICAL alerts (>95%), and generate alert message and denote severity (🚨, 🔴, 🟡).

Notification Deployment

The final, polished message from the LLM is sent immediately to the relevant team via the Gmail API, completing the loop.

⁠

Impact & Achievements (The Bottom Line)

These points show the value you bring, framed around efficiency and innovation.

100% Automation: Achieved full autonomy from data ingestion to notification, eliminating the need for manual monitoring or template editing.

LLM Integration Success: Demonstrated skill in integrating Generative AI into a core business process, replacing rigid logic with dynamic, natural-language analysis.

Enhanced Response Clarity: The system ensures that alerts are highly informative, reducing time required for engineers to understand and action the issue.

Robustness & Scalability: Successfully implemented a production-ready PostgreSQL setup and mastered complex Google OAuth 2.0 for reliable external service integration, proving the system is robust and ready to scale.