I use tags a lot to filter my publications. Right now my list of keywords has grown to levels that are a bit unwieldy. I love the feature that lets you click on a keyword and make all unrelated keywords disappear. This is an extremely useful mechanism to browse my library. But, I would like to change it to something more graphically engaging and interactive.

Ideally it would be a cloud of keywords, linked by their co-occurrence association strength, with each node clickable, and a memory to only display publications that have tags that have been clicked. This would essentially a new way to browse my references in an interactive, graphical manner.

To get there, I exported my keywords collection, and converted it to json using a python script (available at https://github.com/Inserm-CRI-LBI/ZoteroKeywordsGraphVisualization.git)

Then I display them using a d3 layer, with a proof of concept started in Observablehq here:
https://observablehq.com/d/070e0515f4a63540

Does anyone out there have similar needs (and better programming skills than me) and ideas about how to pursue this as an integrated Zotero plugin ?
https://s3.amazonaws.com/zotero.org/images/forums/u12818134/ytglu4g6f2uhdb0fa5vx.png