2017 December Webinar: Best Practices - Machine Learning & Data Science with the PI System

Version 2

    Please download the attachment "OSIsoft Q&A Dec Webinar.pdf" below for the complete document.


    Q: How scalable is AF in terms of the number elements / PI points? 

    A: We have some customers with several million Elements with millions of Attributes pointing to PI tags.  Make sure that you put the elements into a hierarchy and not all in a flat hierarchy, i.e. at the same level.

    Q: In addition with R and MatLab - is there any intention to integrate Pi with Python in the near future?

    A: Yes it is very likely to be after the MatLab integration.  We are looking at integrating both R and Python, but do not have a specific timeline yet.

    Q: Will R and MatLab functions applicable to any elements or analysis variables?

    A:  Attributes from the Elements will be passed to the functions created in R or MatLab as parameters, then R or MatLab runtime would execute the function, and return a result into the Variable in the Asset Analytic expression

    Q: So doesn’t this just make you use extra pi tags to do calculations that you could do in excel with the single initial tag? (In relation to the demo)

    A:  This was just a simple example, but there is a difference in performing it in the analytics and saving in Pi tags over doing it in Excel.  If the required results are only needed in a report and no other tool, then doing it in Excel makes perfect sense. But, if other analytics/applications then I would do it like I showed in the example.

    Q: Is Falkonry part of Azure?

    A:  No, they are a partner of OSIsoft.  You can find their contact information on our partner website

    Q: How much does the complexity level and legitimate value vary is your process is not static or consistent: research loads, machine config., environment

    A:  Please contact me, and we can go over this with some of your use cases. If the process is not static in terms of configuration, it does make it more complex.


    Watch On-Demand version of this webinar

    Download Slide Decks in PDF

    001714_MASTER Machine Learning FINAL.pptx - PowerPoint_2017-12-14_1356.png


    Additional resources:


    • Migrating to the Latest and Greatest Asset Analytics



    Ales Soudek

    Ales Soudek's Blog