Add AI model to environment monitor lab

[jimbobbennett]

The environment monitor could do with an additional step showing how to build an AI model using Azure Machine Learning Studio to do something more with the data.

This model would need to be trained on some kind of existing data set - such as a mocked up data set, and use this to make predictions.

For example, the sound level could be noise from a machine, and if this increases along with temperature it means the machine is about to fail. A dummy data set of temperature/noise to chance of failure could be used to train a model, and then use this from Stream Analytics to report the chance of failure based on live data from the Pi.