Analysis
The Analysis panel computes the network measures that drive the visual encoding: centralities, communities, directed-structure metrics, global statistics, and topological invariants. Every result is stored as a named mesh attribute.
The parent panel requires a graph to be loaded.
Subpanel — Centrality Metrics
Node-importance measures for undirected graphs.
| Control | Description |
|---|---|
| Method | Degree, betweenness, closeness, or eigenvector centrality. |
| Calculate (Node Importance) | Computes the chosen centrality and stores it as an attribute. |
| Calculate (Local Clustering) | Computes the local clustering coefficient per node. |
Subpanel — Community Detection
Partitions the graph into communities using the SurpriseMe algorithms (pysurprise).
| Algorithm | Reference |
|---|---|
| CPM — Clique Percolation Method | Traag et al., Phys. Rev. E 84 (2011) |
| Infomap — map-equation framework | Rosvall & Bergstrom, PNAS 105 (2008) |
| RB — Reichardt–Bornholdt Potts model | Phys. Rev. E 74 (2006) |
| RN — resolution-free Potts model | Ronhovde & Nussinov, Phys. Rev. E 81 (2010) |
| RNSC — Restricted Neighbourhood Search | King et al., Bioinformatics 20 (2004) |
| SCluster — hierarchical clustering (Jerarca) | Aldecoa & Marín, PLoS ONE 5 (2010) |
| UVCluster — iterative cluster analysis (Jerarca) | Arnau et al., Bioinformatics 21 (2005) |
| Control | Description |
|---|---|
| Clustering Algorithm | One of the seven algorithms above. |
| Detect Communities | Runs detection and assigns a community id per node. The Surprise quality metric is computed automatically. |
Subpanel — Directed Analysis
Note
When it appears: only for directed graphs. It groups three child subpanels.
Directed Centrality
| Control | Description |
|---|---|
| Method | PageRank (web importance), hub score (points to many authorities), authority score (pointed to by many hubs), in-degree (popularity), out-degree (influence), or Katz centrality. |
| Calculate | Computes the chosen directed centrality. |
Structure Analysis
| Control | Description |
|---|---|
| Detect Patterns | Classifies the graph: DAG/acyclic, presence and count of cycles, strong connectivity, and the number of strongly connected components. |
| Find SCCs | Labels strongly connected components and reports the component count and largest component size. |
Flow Analysis
| Control | Description |
|---|---|
| Analyze Flow | Classifies node roles into sources, sinks, and intermediaries and reports their counts. |
| Mode | Flow animation mode: Discrete (binary 0→1 switching) or Continuous (smooth gradient wave). |
| Speed / Smoothness / Loop | Animation pacing, gradient smoothness (continuous), and looping. When looping, the panel reports how many cycles fit in the timeline. |
| Create Flow Animation | Propagates activation from source nodes through the directed topology, writing flow_activation and flow_distance attributes. |
Subpanel — Statistical Analysis
Global graph metrics.
| Control | Description |
|---|---|
| Calculate Statistics | Computes density, global clustering, diameter, average path length, assortativity, and the degree distribution (mean, median, std, range). Results are displayed inline. |
Subpanel — Topological Analysis
Switches between analysis modes with a mode selector at the top. The Surface Embedding child subpanel is shown in Surface mode.
Surface Embedding
Note
When it appears: when the topology mode is set to Surface.
| Control | Description |
|---|---|
| Embedding Surface | The target surface type (e.g. Plane). |
| Check Planarity | Boyer–Myrvold planarity test; for non-planar graphs reports the Kuratowski subdivision type (\(K_5\) or \(K_{3,3}\)). |
| Calculate Genus | Reports the genus lower bound and the Euler characteristic (\(V\), \(E\), \(F\), \(\chi\)). |
| Compute Faces | For planar graphs, detects faces and writes the face_id attribute. |
| Create Dual Graph | Builds the geometric dual \(G^*\) (vertices = face centroids, edges between adjacent faces); the dual can be toggled or removed. |
| Compute Embedding | Computes a crossing-free straight-line drawing (Chrobak–Payne for planar inputs) and generates a surface mesh. |
| Validate Crossings | Verifies the drawing has no edge crossings and reports any found. |
Results Summary
Note
When it appears: once any topology analysis has run.
A read-only summary of all topology properties: planarity, genus bound, face count, cycles, linked pairs, intrinsic linking, and the Euler characteristic.
Typical workflow
- For an undirected graph, compute a centrality and detect communities; for a directed graph, use Directed Analysis.
- Run Statistical Analysis for global descriptors.
- Use Topological Analysis for planarity, genus, and crossing-free embeddings.
- Map the resulting attributes to colour/size in the Visualization panel or stratify them with the Network Splitter 3D.