Nowadays, the increase on the prevalence of neurological diseases has launched the development of new methods for their diagnosis and treatment. One of the most common diseases is the Stroke, whose focus on behalf of researchers has mostly been to identify how motor deficits affect stroke survivors in a different way. In Stroke Rehabilitation, the Kinect Device has recently been playing an important role, whose components make it a noninvasive and inexpensive device capable of tracking human motion in 3D. This work focuses on the analysis of activities of daily living recorded by Kinect, and the construction of 3D histograms based on joint angles from the elbow and shoulder joints as models used for the analysis. To do so, a new MATLAB function was developed to draw a bubble plot representing probability of finding combinations of the three joint angles during each task in 3D. Also, correlations between histograms on the set of daily activities were computed to describe their relationships. These values along with the visual interpretation of the histograms will help the reader to understand the conclusions of this project. The results showed that, although Kinect is not the most accurate device, it could be used to construct models representing behaviors of different groups of people performing tasks. These models ended up being explanatory of differences between tasks, tasks and free exploration, and they were also helpful to understand deficiencies presented in stroke survivors, which means this could even be developed as a diagnostic tool in the future.