🏗️ Structural Health Monitoring — Automated Weekly Report (n8n + OpenRouter)

This project automates the generation of a weekly structural health report
using ThingSpeak IoT sensor data, processed and analyzed through n8n and an LLM model (OpenRouter / OpenAI).
It combines IoT data retrieval, preprocessing, AI-based analytics, and automatic reporting — all in one automated workflow.

---

🔁 Workflow Overview

Automation Flow:

1. Cron Trigger – Runs automatically every Monday at 08:00 (Europe/Paris time).
2. HTTP Request – Fetches the last 7 days of data from the ThingSpeak API.
3. Code Nodes – Clean and organize JSON, then compute weekly statistics (min, max, average).
4. Basic LLM Chain (OpenRouter / OpenAI) – Uses an AI model (e.g., GPT-4.1-mini) to generate a detailed, structured technical report.
5. Email Node (Gmail) – Sends the final report automatically to the engineering or research team.

---

🧠 Project Goal

This workflow is designed for structural health monitoring (SHM) of small structures equipped with low-cost IoT sensors.
The goal is to:

• Automate weekly data analysis.
• Detect potential drifts or anomalies.
• Correlate temperature with inclination (X/Y).
• Produce professional reports for engineers or researchers.

---

🧩 Repository Structure

├── workflow/
│   └── weekly_sensor_workflow.json     # Exported n8n workflow
├── docs/
│   ├── architecture.png                # System diagram or workflow screenshot
│   └── data_flow.md                    # Step-by-step data transformation description
├── examples/
│   └── sample_output_report.md         # Example AI-generated weekly report
└── README.md

---

⚙️ Requirements

Before running this project, ensure you have:

• n8n (v1.50+ recommended)
• ThingSpeak Channel Read API Key
  (for accessing the IoT data)
• OpenRouter or OpenAI API Key
  (for AI report generation)
• Gmail or SMTP credentials
  (for sending reports by email)

---

🚀 Setup Instructions

1️⃣ Clone the Repository

git clone https://github.com/sajjadhekmat/weekly-sensor-report.git
cd weekly-sensor-monitoring

2️⃣ Open n8n

• Launch n8n via your browser or local instance.
• Go to Workflows → Import from File.
• Select the file:
  workflow/weekly_sensor_workflow.json

3️⃣ Configure Credentials

• Open the imported workflow.
• Add credentials for:
  • OpenRouter/OpenAI API key
  • Gmail (or SMTP) for email
• Update the ThingSpeak Channel API Key in the HTTP Request node (if needed).

4️⃣ Test the Workflow

• Click Execute Workflow to test it manually.
• Verify that:
  • Data is retrieved correctly.
  • The LLM generates a readable technical report.
  • The email is received successfully.

5️⃣ Enable Weekly Scheduling

• Open the Cron Node.
• Set to trigger automatically every week (e.g., Monday 08:00).
• Save and enable the workflow.

---

📊 Data Flow Summary

Step	Node Name	Function
1	Cron Schedule	Triggers the workflow weekly
2	HTTP Request	Fetches 7 days of sensor data (Temperature, Inclination X/Y)
3	Code (Clean JSON)	Cleans data and removes nulls or invalid readings
4	Code (Calculate Stats)	Calculates min, max, and average for each variable
5	Basic LLM Chain	Sends processed JSON to AI for report generation
6	OpenAI/OpenRouter Model	Generates a markdown-formatted technical report
7	Email Node	Sends report to defined recipient(s) automatically

---

🔒 Security Notes

• Never commit API keys or passwords in your workflow exports.
• Use n8n Credentials Manager to securely store:
  • OpenRouter/OpenAI API keys
  • ThingSpeak tokens
  • Gmail/SMTP credentials
• Check that .gitignore excludes files like credentials.json.

---

📸 Example Output

Generated AI Report Example:

### 1) Summary
- **Period Analyzed:** From 2025-11-04T00:07:24Z to 2025-11-05T14:52:06Z (approximately 1.6 days due to limited data provided).
- **Typical Temperature Range:** 14.5°C to 20.5°C.
- **Typical Inclination Ranges:**
  - Inclination X: approx. 0.8° to 1.18°.
  - Inclination Y: approx. 0.59° to 0.70°.
- **Key Conclusions:**
  - Temperature data shows a gradual variation over the scanning period, typical for environmental changes.
  - Inclination measurements reveal smooth variations with temperature, indicating expected structural thermal response without abrupt anomalies.
  - No sensor malfunction or abrupt data gaps detected, suggesting reliable sensor performance.

---

👷 Author

Sajjad Hekmat
Data Scientist/PhD Researcher@UniLisboa Automation with n8n + IoT + LLMs
📧 sajjad.hekmat1995@gmail.com

---

🧾 License

This project is released under the MIT License.
You are free to use, modify, and distribute it with attribution.

---

⭐ Acknowledgements

• ThingSpeak IoT Platform
• n8n.io
• OpenRouter API
• OpenAI API

---

📘 If this project helps you automate your sensor analysis, please star the repo and share it!