Node metrics?

[mick-d]

Great package!

When looking at the network metrics (master/pynets/netstats.py), I could only see global metrics (i.e. one scalar for the whole network). How would one use PyNets code to get local metrics, i.e. at node level with one metric for each node, such as the ones below?

• degree
• global efficiency (efficiency of a node i with all other nodes)
• local efficiency (efficiency of a node i with its neighbors)
• betweenness centrality bi (Influence of a node i over the information flow between all other nodes)
• clustering coefficient (fraction of the neighbors of a node i that are also neighbors of each other)

[dPys]

Some node metrics are already available :) You can open the pickle file 997_SomMotA_net_mets_sps_cov_0.95 for the example subject (997 master/tests/examples/997) to see what this looks like. PyNets calculates eigenvec. centrality, betweenness centrality, communicability centrality, and rich club coefficient (See lines 672-799 of netstats.py). You can expect to see more nodal metrics including the ones you noted in the next version (i.e. the formal "beta" version) to be released in a few weeks...

Thanks for your suggestions!

[dPys]

Hey @mick-d

See development branch.

[mick-d]

Awesome, thanks!