Force directed graph using traceroute data as edge weights? #137

New Issue

2026-10-04T23:02:04+02:00

adam commented

2026-10-04 23:02:04 +02:00

Rather than using geographic layout for nodes, I'd love a view that shows a force directed layout, using signal strength from traces.

Here's a prototype I was trying with just local traceroute data and https://networkx.org/en/:

Rather than using geographic layout for nodes, I'd love a view that shows a force directed layout, using signal strength from traces. Here's a prototype I was trying with just local traceroute data and https://networkx.org/en/: <img width="2206" height="1304" alt="Image" src="https://github.com/user-attachments/assets/5ce56bb5-cb73-4683-8c80-dc79e6141f78" />

adam commented

2026-10-04 23:02:05 +02:00

I don't think I understand force directed layout? Could you explain?

I don't think I understand force directed layout? Could you explain?

adam commented

2026-10-04 23:02:06 +02:00

If nodes are "closer" together in radio space (higher dB), they'd be plotted closer together, whereas nodes that are further apart would be pushed apart. This should correlate with physical/geographical layout, but may be different at times due to propagation/antennas/etc.

If nodes are "closer" together in radio space (higher dB), they'd be plotted closer together, whereas nodes that are further apart would be pushed apart. This should correlate with physical/geographical layout, but may be different at times due to propagation/antennas/etc.

adam commented

2026-10-04 23:02:07 +02:00

Cool idea. Are there any specific algorithms you would recommend?

Cool idea. Are there any specific algorithms you would recommend?

adam commented

2026-10-04 23:02:07 +02:00

Looking at your stack, mostly ruby, not sure. I don't have direct experience with a Ruby implementation of weighted networks like this.

The above example was https://networkx.org/en/ in Python.

Looks like there's a rubygem wrapper around GraphViz which I've used before: https://rubygems.org/gems/ruby-graphviz/versions/1.2.5

GraphViz can make diagrams like these:

Weight or length would be set to the dBi:

On the JS side, d3 might be closest, or but not sure if you want a JS dependency.

Looking at your stack, mostly ruby, not sure. I don't have direct experience with a Ruby implementation of weighted networks like this. The above example was https://networkx.org/en/ in Python. Looks like there's a rubygem wrapper around GraphViz which I've used before: https://rubygems.org/gems/ruby-graphviz/versions/1.2.5 GraphViz can make diagrams like these: - https://graphviz.org/Gallery/twopi/twopi2.html - https://graphviz.org/Gallery/neato/softmaint.html Weight or length would be set to the dBi: - https://graphviz.org/docs/attrs/weight/ - https://graphviz.org/docs/attrs/len/ On the JS side, [d3 might be closest](https://observablehq.com/@d3/disjoint-force-directed-graph/2), or but not sure if you want a JS dependency.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: adam/potato-mesh#137