From the last decade action recognition is most active research area in computer vision field, which helps us to do the work by training the machine in any field instead of human which improves the time being and human effort and cost efficient. The most used technique’s in the activity recognition are deep learning, optical flow in the advancement in technology several tools are developed by respectable firms for more enhancement of the perception of human behavior, several of them Kinect, 3d data visualizer which data gathering from such systems and data analysis vary from one another, in this report we use Kinect dependent 3 data streams which has 3d view and gathers data in 3 forms of skeletal, RGB and Depth.