How do YOU use Treemaps?


(eastgate) #1

A support query arrived this morning with a stunning example of a treemap – one of Tinderbox’s least-appreciated views. Do you use treemaps? How and when?

Here’s a treemap view of Those Trojan Girls. I’m using the new Ringside Condensed font here, which works well for this demanding chore. One of the times when I could really use a bigger screen!


(Greg Korgeski) #2

Good question, Mark. It got me thinking and doing some exploring. Yes, they are underused, by me anyway. I think that over time a very few uses have come up for me, one I recall being that in some large documents, treemaps help me recognize when the structure of my document has gotten unruly or disorganized with accumulated stuff. I think I did some restructuring of some key things which was helpful.

I like that one can zoom in or out of the treemap views, for closer inspection of these “cities of notes”.

If I had one change, I’d like to be able to tell somehow (in an unobtrusive way) which notes have text in them. You have to zoom in a lot sometimes to see note titles if you have lots of layers but that’s less of a bother for me; the ability to find what has stuff in the note is probably more useful.


(eastgate) #3

Use $WordCount for the Treemap’s color expression – that will make the color of the treemap entry depend on the length of the text.

(See the ⓘ button on the tab to edit the color expression; see the Getting Started pdfs for more background on their use.)


(Greg Korgeski) #4

Thanks, Mark. I see how it works. Can’t find anything in the pdfs (Getting Started with Tinderbox or the Recent Changes/What’s New sections on color expression, though. This helped:

http://www.acrobatfaq.com/atbref6/index/Dialogs/TreemapPropertiespop-ove.html


(eastgate) #5

Actions and Dashboards, p. 63ff


(Greg Korgeski) #6

Thanks! It does enhance the treemaps to use this.

Just to clarify, under Window, there is a “Getting started with Tinderbox” guide and I searched all over that before realizing that you were referring to another guide below it called “actions and dashboards.” I count about three or four built-in guides along with the many other guides including two forums and Mark A’s voluminous documentation. It can take a while to find stuff.


(Greg Korgeski) #7

A bit more thinking about treemaps. I almost wish (actually, not “almost”) they were called something other than “tree- maps.” The name no doubt has an important history in the realm of this branch of programming or visual info display or whatnot. But a “tree” conjures the mental image of – well, a tree. The Tinderbox maps tend to conjure up something more like, well, just “maps.” Or maybe “city maps,” and in fact a city map is what it most resembles to me. I wonder if the visuals are a bit off with the name and that might lead people to “forget to remember” (a hypnosis term) that they are there.

But that might be a loss. Because they can become a bit addictive as a Tinderbox document grows. One reason for that is that they have some potential (not entirely realized yet) to help with something that I often think is a missing element in Tinderbox. Like other users, I kind of live both in several Tinderbox docs and also in both DEVONthink and in TheBrain. I have many thousands of notes in each of these. But I have tended to see Tinderbox more as something akin to using Google, in that while I often know or recall that yes, there is “something in there somewhere” bearing on a topic, it is often not readily visible. The big structure is hard to see in big documents. Yes, you can make some stunning map views or even chart views which give you some clues to connecting notes (and there are other tricks like finding similar notes, which happens in both Tinderbox and DEVONthink). But at times the various sorts of views obtainable in TheBrain can at least _seem to_be a bit more compelling. These are admittedly still usually partial views of a small part of a universe of notes, so the sense of having more “visible real-estate” in TheBrain may be illusory – it’s often more like playing a video game wherein clicking on a thought is like opening a new door that reveals six or sixteen other doors you otherwise did not see (but you then can’t see where you just were.) But I still often find that making relationship connections as I add things can give me at least a feeling that I’m “seeing more of what’s there,” and also that I am actually learning more, making more mental connections.

Treemaps might offer the same benefits. Maybe even moreso. But access somehow to more information – links, notes’ content, is not often present. I suppose one could structure TBX docs to facilitate this (clear container labels and good organization can really help, for instance.) Probably there is already more one can do with them to advance toward that ideal. The universe in a glance – a humble enough aspiration!


(Fernando Rodriguez) #8

I wasn’t even aware of its existence. Interesting!


(Greg Korgeski) #9

As best I recall, Treemaps may have vanished for a bit in a recent version. I’ve been recently reviewing the latest download I can find of the Manual and noticed that, the Treemap view is not listed on the list of possible views, on page 12. They are covered on p. 50 however.

(Edited because I can’t believe I actually somehow wrote “lo and behold…” That was stupid. Sorry.)


(eastgate) #10

Treemap was omitted in Tinderbox 6.0 and returned in Tinderbox 6.3. The omission in the list in the manual section about Multiple Views, Multiple Perspectives will be fixed in the next release.


(jmm) #11

How could I color the notes in a Treemap according to the prototypes assigned to them?

Incidentally, when I click on a note in the treemap view, it doesn’t show in the text pane on the left. However, it can be opened in a text window, and then the text pane switches to that note. I suspect this is not the expected behaviour.


(Rick) #12

I like Treemaps a lot, though currently I’m having problems with them.
In general, I use them with quotations from sources (created in Sente and exported to DevonThink via the Finder, and thence to Tinderbox). I find they work really well when I have given the notes good titles rather than the default ones created by Sente from the first several words of the quotations, and if the text is quite small to enable display of longish titles. I’m able to look at the relevant part of the document and get a good grasp of the content of the source in quite a short space of time, while drilling down into details as necessary by selecting one of the notes.

Though I’m very happy with them (in principle) as they are now, I’m sure that understanding them better would enable me to make them even more useful to me.


(mark seddon) #13

I have created a large treemap that describes a business capability model. It has upto 4 levels of heirarchy if you look at the outline view. What would be useful is to “filter” the view to make it more readable. e.g. make heirachy 3 or 4 invisible. This would aid legibility. Is this possible?


(Mark Anderson) #14

I don’t believe there is functionality to ‘hide’ lower levels. I guess that would be a feature request. as it would need some mechanism to tell Tinderbox how many (outline) levels down to show.

You can ‘hide’ upper levels by simply drilling down into the view so that one of the lower level items becomes the root of the current view.


(eastgate) #15

Rather than filtering out the lower levels, you might want to make the treemap window larger in order to make the map more clear.

But, if you do want to deemphasize interior nodes, you might use the treetop’s Color Expression to lighten notes at greater $OutlineDepth.


(mark seddon) #16

Thanks for the response. Unfortunately it doesnt really help as the map is large and illegible withe the 4 levels of heirarchy. I think I will try exporting to opml and use my mindmap software to filter the levels


(Andreas Grimm) #17

Since I guess it’s not Tinderbox’s Chart-View (which is rather similar to an Mindmap-Appearance) your, @sammas, are looking for - could you elaborate on what you try to achieve that couldn’t already be achieved with the suggestion @mwra made in the second paragraph of his suggestion above?


(Mark Anderson) #18

If I understand correctly, the desire is to limit the ‘depth’ to which the treemap is drawn resulting in fewer, but larger, treemap partitions.


(Andreas Grimm) #19

ok. But shouldn’t exactly that be achievable by using the drilling-method you just mentioned above @mwra?


BTW 01: How do I link to a certain post within a thread? Clicking on the chain in order to copy the link to the clipboard did not work for me).

BTW 02: @eastgate While taking a look in this treemap-discussion I realized that clicking on a note in the (left) View-Pane does not reveal the text/content in the (right) Text-Pane. Just after double-clicking and then again moving one level (back) up displays the text/content (in the text-pane) of the selected note (in the view-pane).


(Mark Anderson) #21

I think Discourse assumes you’ll click the link icon shown bottom right of each post:

link-icon1

Anyway, it’s what I always (successfully) use.

Even if the view is hoisted, the root note of the view can always be selected. You might need to select it, i.e. if nothing is selected, simply click on the upper-most part of the treemap view.

However, I think I can replicate this. An exception to the above appears to be if you double-click a leaf item, i.e. one with no child, and hoist it. The item is shown in the view pane’s breadcrumb bar but both view pane and text pane have no content (and no selection).