Interesting tweet from <@UEH5JQS0J> on trying to o...
# thinking-together
r
Interesting tweet from @Marcin Ignac on trying to organise and navigate the hairball graphs that can emerge from note-taking apps. https://twitter.com/marcinignac/status/1282625655725858817
👆 1
👍 3
o
This is nice, I’ve always loved force directed graphs, but it’s always felt obvious that we should never consider them the universal basis for our connected data visualisations that they’re sometimes touted to be. What they do provide though, is a reasonable and completely universal fallback visualisation. If all else fails, do some basic force directed layout, but there are always going to be better approaches to pulling out the essential information behind the data
💡 1
And importantly, there are often many visualisations that reveal many different aspects of the data
Lastly (rambling I know, sorry!) it’s worth noting that we typically see graph data for very simple things, just because there are millions of vertices doesn’t actually make it complex (people, wiki pages, airports, etc) but it is just as reasonable to embed many overlapping domains in the same graph.
💯 3
r
Agreed. I suppose this is effectively summarising the graph into a smaller, more legible, 5 node graph (in which you can see the colour-coded sub-nodes)
w
Huh. I try to steer clear of basic force directed layouts because, simply, you don't know which way is up. I mean humans are naturally oriented creatures as are most of the things we usually interact with. Beyond that, I've had shocking success by simulating "breezes" passing through a network. Since it's how trees work, and maybe since we're primates, the slight perturbatious somehow make the the connections pop out.
💡 2
o
@wtaysom have you got an example for that? I love the idea of a ‘breeze’ as it’s a very natural metaphor. It seems it also fits into a more generalised idea of using forces over time to elucidate things that aren’t clear with a static graph. Of course it’s possible to layer more semantics onto graphs such as animating data flowing through edges, or coloring nodes, etc. But I love how far you can go with just nodes and edges and some movement.
1
w
@Orion Reed nothing to show at the moment, alas.