Drop betweenness centrality from facebook notebook.

[rossbar]

This PR is a proposal to drop the betweenness centrality portion of the analysis from the exploratory facebook notebook.

The motivation for doing so is to strike a better balance with respect to computation time. The betweenness centrality computation takes >60s and is more than 50% of the computation to run the entire notebook. This makes it a bottleneck when going through the tutorail interactively.

[eriknw]

Sounds like running BC for this could be a great demo for using accelerated backends!