Jerome Lefebvre

Finding correlation between attributes using PI Vision (an experiment)

Blog Post created by Jerome Lefebvre on Jul 4, 2017

This is simply an experiment, I do not recommend anybody to actually do the below other for entertainment.


I was thinking this morning about the PI Vision Hackathon's entry on showing how to use a custom symbol to show correlation between various symbols.

PI Visualization Hackathon - Magion Scatterplot Matrix


In PI Vision 2017, there is now an XY-plot that comes as a CTP, thus it is possible to compute correlation between between 2 attributes. The limitation is now if you want to compare multiple attributes at the same time.

XY-plot cannot currently be part of a collection and confiscating many at the same time can take quite some time. Doing so manually can take quite a long time. What one can do instead (as an experiment!) is automatically create new displays base on the list of tags you want to compare.


This is what I did in the sample below:


What this sample does is generate the underlying code for a displays as follows:

I have a list of 4 attributes センサ1,センサ2、センサ3、センサ4 that I want to compare to each other. Thus, I need to create a xy-plots that compare 1 to 2, 1 to 3, 1 to 4, 2 to 3, 2 to 4 and 3 to 4.

The code does this automatically. Similarly to compare 5 attributes to each other.. I need to zoom out at this point.



The code is nothing special, as all it does is manipulate a JSON object.


The braves can take a look at the code in the codepen above.


The only benefit of this approach over the one described in the hackathon is that if everything goes well and you didn't break PI Vision by using internal methods the resulting display uses only standard objects.