Thank you all for attending the PI Integrator for Business Analytics webinar held on Tuesday, 10th of May 2016 (GMT+10). You can see the webinar recording via the link below.
This post is created to address those questions raised during the webinar and also any follow-up questions you may have regarding the PI Integrator for Business Analytics.
Feel free to let your curiosity run wild. Please also let me know if I have missed any questions.
1. I'm guessing the ease of the BI views will be based on the structure of the AF environment. Are there guidelines on how the structure should be done?
In a nutshell, PI Integrator works best with AF structures that utilises the "template", "category" and "Event Frames" features of AF. See webinar recording between 00:21:34 and 00:22:34 for explanation of AF "templates" and attribute categories.
Categorise AF attributes will help provide a cleaner end user experience to find the attribute of interest.
When uncategorised, user’s attribute names appear as it is in the PI Integrator for Business Analytics.
When user select “Group by Category”, you can quickly navigate the attribute of interest in its designated category. In this case “Operational Data”.
The explanation on the importance of PI Event Frames utilisation can be found during the webinar. See question 2.
2. How do we use Event Frames as a filter?
Watch the webinar recording between 00:29:36 and 00:30:45 for demonstration. Please contact us if you wish to see more.
3. Is there any possible way how we can do the reporting on SAP BI tool like SAP BusinessObjects?
There are several ways for SAP BusinessObjects to utilise the data published by PI Integrator for Business Analytics. It mainly depends on the BusinessObjects version you have.
According to OSIsoft KB article KB00204, SAP BusinessObjects XI and 6.0 are incompatible with PI ODBC. Hence ways to utilise the data published by PI Integrator for Business Analytics are:
- Use BusinessObjects 5.x or earlier with PI ODBC
- Use SQL Server as the bridge between BusinessObjects and PI Servers. (PI Servers ==> PI Integrator 4 BA ==> SQL Server ==> BusinessObjects)
- Use Microsoft Access as the bridge between BusinessObjects and PI Servers (PI Server ==> PI Integrator 4 BA ==> MS Access ==> BusinessObjects)
4. How long does it take to design the dashboard, let's say for 1 year worth of data view? Can we also see details if we click on graphs for individual drills for instance?
The time it takes to build a dashboard with 1 year worth of data would be similar to what you saw in the webinar between 00:35:00 and 00:40:12. The dashboard shown in the webinar consist of 1 year worth of data across 6 assets/rigs. The dataset size is approximate 3 million rows over 18 columns = 54 million events published into Tibco Spotfire.
You can see details of individual drills if you click on it as shown below:
5. What happens if the event frame is changed in PI?
The PI Integrator for Business Analytics monitors changes made on the AF server and the databases use by the views configured by the Integrator. When anything is added, deleted or renamed, it updates any relevant PI views. For example, when a new event frame is created that matches the asset shape of an existing PI View, that event frame is added to the PI View. Data for the new event frame will be backfilled to the PI View start time during the next run. When an event frame is deleted, no new data will be written for that event frame. The existing data will not be deleted.
6. How the data cleansing is done in PI Integrator for Business Analytics?
You can see the data cleansing feature of the PI Integrator during the webinar recording between 00:27:37 and 00:29:36.
Please contact us if you wish to see more.
7. Does this tool have predicative analytics capabilities?
The PI Integrator for Business Analytics is designed to help users to transform time-series data from PI system into a data set that is easily consumed by Business Intelligence (BI) tools. It has the capability to publish future timestamp data as well as historical data from the PI System to other business systems such as MS Machine Learning and SAP HANA. These are the conventional platforms where predictive analytics are performed by business users. Having said this, PI users can also perform basic predicative analytics by utilising AF analytics within the PI Server suite, where calculations with future timestamps can be easily configured.
Here are some reference materials to help you to understand PI future data/analytics capabilities:
Short Youtube videos on Future Data: https://www.youtube.com/watch?v=0MglijxFCcM
Customers Talk about Future Data in PI Server 2015: https://www.youtube.com/watch?v=kZzAj1u3Yqg
Documentation on Future Data: https://techsupport.osisoft.com/Documentation/PI-AF-SDK/html/78e1ff0f-f0fa-4843-9156-e3000fd7586b.htm
Discussion thread regarding “Future/Predicative” data using PI Server 2015: https://pisquare.osisoft.com/thread/10908
News Release on PI Future Data: http://www.reuters.com/article/ca-osisoft-idUSnBw075005a+100+BSW20150707
Please contact us if you wish to know more.