Cold Chain Monitoring Dataset: Temperature, Humidity and Vibration Data for TinyML Classification

This dataset was collected as part of a Master's Thesis project focused on IoT and AI-based cold chain monitoring using ESP32CAM microcontrollers and TinyML models trained on Edge Impulse.

It contains two subsets: temperature and humidity readings (DHT22 sensor) under correct and anomalous conditions, and accelerometer readings (MPU6050) under three transport scenarios: normal, dangerous vibration and sudden shock.

The data was used to train lightweight neural network classification models deployed directly on the ESP32 CAM, enabling real time anomaly detection without cloud dependency.

This dataset is associated with the work: "A Low-Cost Edge AI Prototype  for Sustainable Cold-Chain Monitoring in Resource-Constrained Settings". 

Associated thesis: "Internet de las Cosas e Inteligencia Artificial Aplicados a la Cadena de Frío" — Master in Automation and Robotics, UPM, 2026.