A2-FlaviaGrey

From cs448b-wiki
Jump to: navigation, search

Body Movement Collected By Kinect

The dataset I chose is the data collected by a Kinect sensor on a sitting person. The dataset includes movement data in 3 axis (X, Y, Z) for 5 body parts: Head, Left Shoulder, Right Shoulder, Left Hand and Right Hand. The first visualization's purpose was to see whether the data points made sense. I filtered the data set, which includes 8 dates to just one date and visualizaded the X and Y axis data for all body parts:

Point Distribution.jpg

It is easy to make out the general movement of the body, and I used the average to confirm and make the visualization cleaner:

Point Average.jpg

The next question was what body part moves the most, and in which direction?

Xyz bodyparts.jpg

Here we can see that the hands move the most, and the axis where the most movement occurs is Z.

I then added back all the dates into the dataset and asked: Which date has the most total movement?

Total per date.jpg

We can see here that the most movement is on April 1st.

My next question was: Which body part moves the most each day? Is any body part consistently the one with the most movement?

Total per bodypart.jpg

The hands are indeed always the ones that move the most, but sometimes it is the right hand the one with the most movement and sometimes the left hand.

I decided then to do the distribution for a single day, to better be able to identify the movement for each day. The following is an example for February 18th.


Day per bodypart.jpg

Finally, I brought all the information I had learned from the previous visualizations into a dashboard. The final question was: Which days was there the most movement, and what were the dominant body parts?

Dashboard.jpg