Embedded machine learning is the process of running machine learning algorithms (including deep learning) on embedded systems, such as microcontrollers and single board computers. These videos come from the Coursera course: Introduction to Embedded Machine Learning. If you would like to take the full course, complete projects, and earn a certificate in embedded machine learning, please go here: In this video, Shawn demonstrates how to collect vibration and motion data (using an accelerometer) from a smartphone and Arduino Nano 33 BLE Sense. This data is logged to a project in Edge Impulse. Additionally, Shawn discusses what makes for a good dataset and how datasets can be divided into training, validation, and test sets. The pitfalls of an unbalanced dataset are discussed along with how datasets can be created to perform inherently unbalanced tasks, like anomaly detection.