Hey Miro-cle team!
Before I get to my request let me just say that I just LOVE that tags exist and that I can use them to organise crazy amounts of data, thank you!
My request is for a more complex/advanced clustering for tags, my use case is qualitative data analysis and my request is for advanced clustering to be able to synthesize data more efficiently. There are three solutions for what I need that I can see, however it’s very possible there are many more 
:
- Multi-tag similarities: at the moment when clustering by tag the same sticky will appear on the named cluster for each of the tags that exist on the board - which is great for some analyses, so definitely would be good to keep. However I would also really like an option to remove sticky duplication - a single sticky appears in a single cluster. That would mean a more complex clustering algo, but it would be great that the clusters could be created based on how many tags the stickies have in common, not just IF they have a single tag in common.
- Layered clusters: This is a different option that I would be really happy, although would also be great to have a way to toggle between stickies getting duplicated/not. After the first round of clustering (with either duplicated/not stickies) it would be great if I could single out one of the clusters and cluster again within it. Is its sort of possible right now, but I would like the new baby-clusters to stay within the boundaries of the first one, so it’s easier to keep track of their hierarchy.
- Hierarchical tags: If the tags could be structured to have layers, this could make it easier for clustering algo how to put them together. Having tags groups OR tags & sub tags would be really useful. For example if I had a tag group called “navigation” by tags within that group could be “hamburger menu”, “home button” etc. And then when clustering there would be obvious layers to cluster by.
I would deliriously happy if either of the three was implemented 
Hope this makes sense, happy to give more detail on my thinking if useful :)
Thanks!
J.
