How should data be structured to produce a Cluster Dendrogram?

31 views
Skip to first unread message

Chris Cuny

unread,
Dec 21, 2019, 4:31:08 AM12/21/19
to RAWGraphs (formerly Raw by DensityDesign)
Hell all,

First time visit to RawGraphs today and my first question on this group.

I am blown away with the quality of the graphs and particularly interested in the Cluster Dendrogram which is the 3rd example in the gallery. It is the "Platforms and actors of the Breton open data" example by Boris Mericksay (I recommend a visit to his site: it's gorgeous). My ultimate reference is John Manoogian III's poster of the Cognitive Bias Codex here

Unfortunately, the documentation does not cover the Cluster or Circular Denrogram and I have no idea how the data should be structured before importing into RawGraphs. I have tried to use the JSON (attached) file containing the same data structure as used for the picture (I also concerted it to Excel (attached) to try it that way via CSV) and I get very confusing results, primarily because I can't structure the data properly (I think). Here is what one run looks like (see bottom of the message).

I have never used JSON and I am newish to Data Visualisation, so I my approach is probably naive, but here are a few questions I am asking myself:

  1. In the beautiful Cogbitive Bias Codew image, you can see that the top level (names) is labeled outside the Dendrogram. It may be that was added manually and that's OK, but I don't want to see that top level labeled inside the Dendrogram and I don't know how to do that.
  2. Maybe the JSON file was edited before the Codex was made (I don't know what was used to make it). In any case, I can't visualise the JSON File. But looking at the Excel file, I find it really confusing anyway. In any case,
  3. The "second Level" of categorisation of data should be the 20 descriptions that can be found on the outer circle with colored dots on the Codex picture. They are the statements that typically start with "We...".. As you can see from my sorry attempt, they're labeled inside the Dendrogram. In the codex picture, this may have been done manually again. In any case, I do want that categorisation to stick, but not labeled as part of the inside of the Dendrogram.

In any case, after playing with mapping dimensions, I am just getting more confused. I just don't know how to get to a sensible starting point that matches the spirit of the Codex picture. I am not worried at this stage about the vusual formatting (except maybe to fin out if there is a way to adapt the font used and/or its size to make the output more legible).

I don't expect anyone to do my work for me (actually, it's not work: I just took my favourite data viz and used it as a model to learn about RawGraphs) but I would really appreciate some pointers in my quest to reproduce the bare bones of the Codex. All tips and suggestions are gratefully acknowledged.

Thank you very much for your comments.

All the best during the holiday season.

Chris

Cog.xls
Cognitive Bias Codex.json
Reply all
Reply to author
Forward
0 new messages