Webinar - Asset Analytics native integration with MATLAB

Video created by rdavin Employee on Mar 22, 2018

    This is the video recording of the PI Developers Club webinar Asset Analytics native integration with MATLAB presented on March 21st, 2018.

     

    Summary:

    In this PI Developers Club webinar, we offer a sneak preview of the AF 2018 release with the highly anticipated feature of native integration between Asset Analytics and MATLAB!  Panelists include Stephen Kwan, AF Product Manager, along with  representatives from MathWorks, Tim Choo and Sachin Nikumbh, both on the MATLAB Production Server team.  Data scientists, developers, system integrators, or anyone interested in analytics is encouraged to attend and ask questions.

     

    Panelist:

    • Stephen Kwan , PI Server Product Manager
    • Tim Choo , MathWorks Inc., Product Manager for MATLAB Production Server (MPS)
    • Sachin Nikumbh, MathWorks Inc, Engineering Manager for MATLAB Production Server

     

    Slides:

     

    MATLAB integration with Asset Analytics, as well as arrays within Analytics, is a very popular topic.  The two-edged sword of such a popular topic is that there is going to be a lot of material covered as well as a lot of questions, and the more material we cover leaves less room for questions!  The recording addresses 15 minutes of questions being addressed, but we did not have enough time to get to them all.  Below are the questions we addressed afterward:

     

    Questions answered after video recording:

     

    Q: How do you plan to utilize the AF External Integration framework?  Is this based on REST calls?  Can it be potentially any REST web service?

    A: Currently it is based on REST calls to MATLAB Production Server.  In theory it can be any REST web service but programmatic access is not available to the public.

     

    Q: In addition to MATLAB, what other technologies are planned for similar integration with AF in the near future?

    A: We’re looking at R and Python.  Customers are encourage to provide us use cases and/or upvote existing User Voice suggestions.

     

    Q: Will PI tags eventually have array storage capabilities?

    A: Not currently in plan.

     

    Q: About PI AF 2018 Asset Analytics: Will it be possible to call custom DLL functions and replace custom data references?

    A: No

     

    Q: Is the option to manage external sources in System Explorer new to the coming update?  Are there any other sources currently supported besides MATLAB Production Server?

    A: Yes, this is new to AF 2018.  No, MATLAB Production Server is the only supported external system with the AF 2018 release.

     

    Q: Are you planning a Microsoft R integration?

    A: Under consideration.  Customers are encourage to provide us use cases and/or upvote existing User Voice suggestions.

     

    Q: In Asset Analytics will the user be able to inspect the contents of arrays?

    A: Yes, we’re working on the UI in PI System Explorer to give you that capability.

     

    Q: Will this work for Table Lookup Data References configured as a time series?

    A: Yes, as input attributes.

     

    Q: Will it be possible to call custom DLL functions (i.e. written in .NET) from PI Asset Analytics for advanced calculations, and thus replace some of the existing PI AF custom data references?

    A: No