It is understood that while breathing is an autonomous activity, it can also be controlled to some degree through conscious effort. Further, breathing is known to be affected by physical and mental conditions. Respiratory illnesses such as influenza and COVID affect our breathing and their early detection can lead to effective treatments. To record and study an individual’s raw breathing in non-clinical settings while also being reliable and unobtrusive requires a special device. One such device is the Spire Tag, a wearable sensor that makes measurements 25 times a second, allowing us to take a close look at the breathing process throughout the day in the wearer’s everyday setting. In this paper, we study impaired respiratory function caused by illness and its effects on the breath shape. To make use of the Spire Tag, we first study the sensor’s linearity, noise, and hysteresis. We then describe Anemoi, a system to process and analyze respiratory function on a breath-by-breath basis. We use Anemoi on data collected through a cohort study of participants that tracked their wellness and illness. Each participant wore a Spire Tag 24/7 yielding over 2.1 million readings per day. For this study we select the data from the subjects during their sleep period. The raw data is processed and segmented into individual breaths. We describe the notable differences in the derived breath features from the days when a subject is healthy and when sick. Results of the classification yields approximately 70% accuracy. Other details will be presented.