Node and Graph Metrics
To help visualize potential relevant structure or important nodes in the data, TempNetViz provides the possibility to compute several metrics. These metrics determine how nodes are sized, colored, and how statistics are displayed.
Available Metrics
| Metric | Description |
|---|---|
| None | No node-specific metric is applied. The graph displays raw edge weights only. |
| Strength | The sum of edge weights connected to a node. A higher strength means the node has stronger or more numerous connections. |
| Betweenness | Measures how often a node appears on the shortest paths between other nodes. Nodes with high betweenness act as important bridges in the network. |
| Closeness | Reflects how close a node is to all other nodes in the network. Nodes with high closeness can reach others quickly (shorter average path length). |
| Eigenvector Centrality | Measures a node’s influence based on the idea that connections to important nodes contribute more than connections to less important ones. |
| PageRank | A variant of eigenvector centrality. It gives higher scores to nodes connected to other highly connected nodes. Originally developed for ranking webpages. |
| Hub Score | In the HITS algorithm, hubs are nodes that point to many authoritative nodes. A high hub score means the node is a good “information provider.” |
| Authority Score | The complement of the hub score in the HITS algorithm. A node has a high authority score if many hubs point to it, making it an important “information source.” |
| Rich-Club | Identifies a subset of nodes (the “rich club”) that are highly connected among themselves. The parameter k controls the minimum degree for inclusion. Nodes outside the club are down-weighted. |
| K-Core | Identifies the largest subgraph where every node has at least k connections within the group. Useful for finding tightly knit groups. |
Notes
- For Rich-Club and K-Core, you must specify a degree parameter
k. - Metrics affect how nodes are displayed:
- Size → proportional to the metric value.
- Color → mapped via the selected colormap.
- In histogram mode, metrics determine the values plotted.
- When None is selected, histograms display edge weights directly.