Hi, just had some ideas for improving graph view in ways that would be beneficial to the community, would like to hear others take on them
A. Ease of use
Size of a vertex should be proportional to the number of vertices linking to it (i.e indegree), would help identify which ideas are most commonly seen in one’s notes (not to mention make them easier to click)
When hovering over a vertex, not only should the title of the node be larger and all the edges colored (as it is today), but only the titles of connected nodes should be visible (i.e. any vertex that isn’t directly connected to the node being hovered over should have its title hidden, they should simply be circles). Makes the graph a fair bit more legible
Would it be possible to optionally hide nodes of a certain type entirely, such as images or daily notes? They can clutter up the graph a bit when I’m trying to look for connections between my standard idea notes
B. Idea Discovery
It should be possible to color the vertices by the folders they’re in, would make it substantially easier to find cross-disciplinary connections
Edges should ideally be directed arrows (if both vertices reference each other, then a double headed arrow works), makes it easier to see which ideas lead where.
The current visualization method seems to be a bit messy with vertices all over the place (not sure if that’s just me though, although something like this, a more "tighter " visualization would be nice: Vis Network | Example Applications | Les miserables). Would it be possible by giving the user access to control the value of the force parameter, perhaps hidden by a down arrow button in the top right of graph mode, clicking on which would reveal a volume dial?
I created my account to suggest hiding daily notes. Maybe the ability to filter the graph by certain tags? You could tag your daily notes and then just filter out that tag, so that it can be toggled on and off.
I would also like to hide template notes, which could be accomplished the same way.
I believe your first bullet point is already Implemented. As long as you keep linking references to the same vertex it will increase in size so that every arc connecting to it can be seen appropriately, thus giving it the look we expect: bigger nodes are important nodes.
But that was precisely my point. I have the exact same result. I was just point out to irhum that the first point on his request was already implemented.
