Visualizing a Multidimensional Data Structure Hierarchy

Question

I am trying to design a system that is capable of displaying the hierarchy of a collection of containers containing many sub nodes to a common user.

The underlying structure is actually a graph, but our software currently displays it as a basic tree hierarchy such as this:

|- Parent1
 |- Child1
  |- GrandChild1
|- Parent2
 |- Child2
  |- Grandchild2

However, the issue is we are in the process of implementing far larger data containers that hold lots of (relatively) smaller containers. Each container holds general information about all the subsequent children. We currently have a data set that is roughly 1/2TB in size consisting of 400+ multidimensional large containers (a 3D collage of sorts) each containing smaller containers (2D images) consisting of containers (attributes) holding collections of arrays that have information about specific features of the underlying data.

In order to allow users to interact with these containers I have been considering the following two options:

• ArcTrees
• BarCode Trees

More so the ArcTrees than the latter, but these are all I could find. In the ArcTrees paper I saw this image:

Where I figured I could make these expand dynamically when the user clicks on them to make a natural visualization similar to the existing tree hierarchy. Is this a reasonable way to do it? Is there a better way I am missing? The underlying structure is a graph, so is this even worth the trouble from a usability perspective?

Also does anyone have examples of software that uses these so I can demonstrate proof of concept to my superiors (I can barely find any mentions of these throughout the internet)?

Thanks in advance for your time and consideration.

PS: Jumping in front of it, I am primarily a back-end engineer who is filling another role at a small company while the front-end engineer is out on leave, so I am not very familiar with the UI territory apologies in advance.

Answer 1

It will depend on whether you need to use this only for visualization or interaction.

Either way, I think your examples will be very difficult to read or interact with 400+ items. So, I believe your best bet is a Radial Tree (and an interactive Radial Tree if you need users to interact).

If this is the case, you can use FuncTree2 by Tokyo Institute of Technology (see the image below).

It's mainly used for biometrics, but it would work for your use case. As you can see in the image, it conveys some information on what will users find on each container, and they can interact on a "per container" basis, expanding and focusing on any of them as needed

EDIT: I found a paper on FuncTree2 at PubMed, which clearly explains the problems it addresses, so I believe you may find it useful.