Network Diagram for showing relationships

[bfoulkes]

I want to be able to show the relationships between two groups of data.  I have seen this as a Network Diagram or a Social Network plot, but JMP can't do this right now.  There is an add-in that will create a plot that is not that interactive, but I would like this as a full plot, where items can be filtered and sizes don't need to be coded, so they can be changed on the fly.

The add-in shows the use case as transactions from one group to another, but this could also be the flow of data from one group to another or a more general social network connections plot.

[MathStatChem]

I think there is a need for network diagrams (i.e. directed graphs, direct acyclic graphs, ...) and graph analysis.  Currently there is the underpinnings of this already developed, the  object (or analysis object, not sure what it is really) Node Graph and the display object Node Graph Display() are undocumented but can be used in scripting to create network diagrams.  

Use cases:

• Visualize and analyze probabilistic decision networks
• Visualize manufacturing batch lot geneologies and do correlation analysis of results at different layers of the geneology graph
• General spatial association analysis (social network analysis, movement through a network, etc...)
• Revamp the "Fishbone Diagram" to use the directed graphs as the drawing tool.  It should allow for interactive creation of these types graphs, with varying orientations/arrangements (e.g. mindmaps, flow charts, fishbone).  

[bfoulkes]

I did come upon this page that shows how to do it with JSL, https://www.jmp.com/support/help/14-2/associative-arrays-in-graph-theory.shtml

[Sarah-Sylvestre]

Hi @bfoulkes, thank you for your suggestion! We have captured your request and will take it under consideration.

[GregMcMahon]

I'd like to throw my support for adding network diagrams into JMP as well.

[BigBook]

The relationships between the objects we study can be expressed in terms of network structures, and large amounts of data are indeed more suitable for network modeling. As a full-featured data analysis tool, JMP is in dire need of complementing network-based data analysis.

Probably the biggest question about whether this feature should be developed is this analysis function has traditionally been largely based on graph theory. It is not a statistical analysis technique.

But it should be noted that the analysis of network structure has introduced a large number of dynamic structures and object randomness under stable structures due to the wide application of computers, so that network structures also have random characteristics and have become typical statistical research objects.

The graph theory tool has actually become the underlying analysis technique.Analysis of dynamics and randomness undoubtedly falls under the umbrella of statistical techniques.

So please develop the analysis function of the dynamic random network. It is very useful!