Method Gallery#
Every method in this package, drawn the same way: the backbone over its original graph, with kept edges in blue and removed edges in faint gray. All figures of a dataset share one layout, so a node sits in the same place in every figure and the methods can be compared at a glance. The figures are generated when the documentation is built, so they always show what the current code returns. Each caption gives the edges and nodes the backbone keeps.
Side by side#
The executed example below runs the statistical and structural methods on Les Misérables and draws them as small multiples, with a bar chart of the edges each one keeps.
One figure per method#
The full-size figure for each method is on its API page, next to the function’s documentation:
Statistical Methods (Les Misérables)
Structural Methods (Les Misérables)
Proximity Methods (Les Misérables)
Hybrid Methods (Les Misérables)
Unweighted Sparsification (unweighted Les Misérables)
Bipartite Methods (Davis Southern Women, including a signed backbone)
Hypergraph Methods (Davis Southern Women as a hypergraph)
Temporal (the synthetic snapshots of Temporal Backbones)
The figures are drawn by scripts/generate_visualizations.py with the
shared figure style in docs/_figures/style.py.