Accelerometer motion classification

Data science demo by Thomas Wood, Fast Data Science Ltd

This model reads in 3D values taken from an accelerometer and classifies a person's motion into 10 different categories such as walking, sitting, falling down, etc. This is a very simple machine learning model to build. To classify motion accurately you need to take into account the absolute values of all three co-ordinates at every moment in time as well as their rates of change.

To test the model please enter some position values in three columns, in format X, Y, Z. I have pre-filled the text box with an example to help you get started. The readings are positions measured in metres, and readings were taken 50 times per second. If you have an accelerometer app on your phone, you could try getting some readings and test how accurate the model is.

What is this kind of model useful for?