Over the Australian summer holidays the three interns, Miwa Teranishi, Ryan Amaudruz, and Thomas Hahn (all from the University of New South Wales, Australia), worked on projects using PI and machine learning tools to investigate the Japanese and Australian electricity markets in OSIsoft's Tokyo, Sydney, and Perth offices.  Even in the short period of time they had to work with us, and a lack of any prior expertise with PI, they were able to develop some interesting models and gain some very interesting insights into these markets.

 

Some of the highlights of the results are:

 

Using PI and Falconry to predict the daily minimum electricity price in Japan

 

Using PI, Microsoft Azure, and only publicly available data to predict net aggregate solar generation in Western Australia, covering an area of 2.6 million square km.

 

 

An overview of their projects, along with some videos describing what they did, and how they did it are available in the blog post below:

 

Three Interns Describe Three Successful Machine Learning Projects

 

Additionally, a version of Miwa's video with Japanese subtitles is available on our OSIsoft Japan youtube channel at the link below:

 

OSIsoft Japan Intern Project 2016 – Power market forecasting with Falkonry - YouTube

 

This is a great example of how the PI system can easily be integrated with other platforms to provide a powerful results in a variety of fields.