📡 IoT Sensor Anomaly Detection using ESP8266, ThingSpeak & MATLAB

This project implements real-time IoT sensing, cloud data logging, and statistical anomaly detection using:

• ESP8266 (NodeMCU)
• DHT11 Sensor (Temperature & Humidity)
• LDR Sensor (Light Intensity)
• ThingSpeak Cloud
• MATLAB Z-Score Analysis

The system uploads live sensor values to ThingSpeak, and a MATLAB script performs anomaly detection on the last 60 minutes of data.

---

🚀 Features

✔ Real-time IoT Cloud Upload

The ESP8266 collects:

• Temperature (°C)
• Humidity (%)
• Light Intensity (LDR value)

And uploads them to ThingSpeak using an API key.

✔ Statistical Anomaly Detection

The MATLAB script analyzes:

• LDR
• Temperature
• Humidity

Using Z-score-based anomaly detection, with thresholds:

Sensor	Threshold (±Z-score)
LDR	1.4
Temperature	2.5
Humidity	2.5

✔ Visualization

A single figure is generated showing:

• LDR Z-score graph
• Temperature Z-score graph
• Humidity Z-score graph

✔ Outlier Classification

The script prints:

• Normal data points
• Anomalous data points
• Total number of anomalies per sensor

---

🔧 Hardware Required

• ESP8266 NodeMCU
• DHT11 Sensor
• LDR + 10kΩ resistor
• Jumper wires
• USB cable

---

🔌 Circuit Connections

Component	Pin on Sensor	ESP8266 (NodeMCU) Pin	Description
DHT11	VCC	3V3	Power supply
	GND	GND	Ground
	DATA	D4 (GPIO2)	Temperature & Humidity data
LDR	One end	A0	Analog light intensity input
	Other end	3V3	Voltage divider
10kΩ Resistor	One end	A0	Forms voltage divider
	Other end	GND	Ground
ESP8266	VIN / USB	USB	Power supply
	GND	GND	Common ground

LDR Voltage Divider

The LDR is connected in a voltage divider configuration with a 10kΩ resistor.

3.3V ── LDR ── A0 ── 10kΩ ── GND

• Bright light → Low resistance → Higher ADC value
• Low light → High resistance → Lower ADC value

🧪 MATLAB Script Overview

The MATLAB script performs:

1. Fetch last 60 minutes of sensor data from ThingSpeak
2. Calculate mean & standard deviation
3. Compute Z-score
4. Classify:
   • Normal values
   • Outlier/anomalies
5. Print summary
6. Generate graphs

---

🧠 Why Z-Score for Anomaly Detection?

Z-score helps detect sudden abnormal deviations by comparing how far a value lies from the statistical mean:

Values outside the threshold → anomalies

---

📊 Output Example

• LDR Anomaly Count: 6
• Temperature Anomaly Count: 2
• Humidity Anomaly Count: 4

Graphs include:

• LDR Z-score plot
• Temperature Z-score plot
• Humidity Z-score plot
  

---

🛠 MicroPython Code (ESP8266)

Includes:

• WiFi connection
• Read DHT11 sensor
• Read LDR through ADC
• Upload to ThingSpeak every 20 seconds

---

📌 Future Enhancements

• Email/SMS alert when anomaly detected
• Live dashboard (React / Node.js)
• Automated ML model for forecasting
• Multi-sensor fusion anomaly detection

---

⭐ Support This Project

If you found this project useful or learned something new from it,
please consider giving it a ⭐ on GitHub — it really helps and motivates me to build more IoT projects like this!

Thank you for your support 🙌

👤 Author

Kritish Mohapatra
Electrical Engineering | IoT | Embedded Systems
GitHub: kritishmohapatra
LinkedIn: Kritish Mohapatra

---

📬 License

This project is open-source under the MIT License.