This paper describes a practical experience with the real-time collection and storage of data from a wearable device used to generate essential datasets for the development of algorithms and artificial intelligence models. The aims of the models are to identify activities of daily living (ADL) movements and to detect falls. The device is equipped with accelerometer, gyroscope, magnetometer, barometer, and microphone sensors, all operating at a frequency of 20 Hz. Initially, significant packet loss was observed during real-time transmission due to distance and data saturation transmitted by Bluetooth Low Energy (BLE). To address this issue, an algorithm was developed to store data in a buffer for 10 s of movement. Subsequently, the data is transmitted to the receiving mobile device in a controlled manner when it is near the wearable device, achieving packet loss-free data transfer. The packet size was specified to analyze its impact on transmission. Two mobile applications were used: one for real-time transmission, and another for collecting data stored in the buffer. Both applications allow data to be saved in .csv files for subsequent visualization. During buffer transmission, a transmission frequency of 10 Hz was implemented on the wearable device to ensure no data loss. This device is designed for data collection in a laboratory setting to obtain raw data that will allow future researchers to create models for identifying ADL movements and falls.