3 ms·
There was a very beautiful idea for solving the same sort of problem back in 1991 called Cone Trees: https://research.tableau.com/sites/default/files/p189-rober
by anigbrowl 4y ago
There was a very beautiful idea for solving the same sort of problem back in 1991 called Cone Trees: https://research.tableau.com/sites/default/files/p189-robertson.pdf https://research.tableau.com/sites/default/files/p189-robert...
Computing speed and rendering power were the barriers for a long time, and those have long fallen. But a larger headache is that navigation gets really unwieldy as the trees grow. I do a lot of stuff with graph visualization (conversational structures, community structures) and you have to do a lot of pruning, filtering, and bundling.
The audience for any given graph is inversely proportional to the number of nodes visible. Once you go past 100-ish nodes layouts start to seem arbitrary and once you go past a few hundred most people's eyes just glaze over. You can make the graph more readable with smart choices about color and node size but it's a constant signal:noise battle.