Text Mining and Visualization

Empirical Analysis of the Stack Overflow Tags Network 285

The results are shown in Figure 12.5. We can confirm, for example, that the presence

of latex in the “R” community is indeed not spurious, as the subject node is strongly

connected with both knitr and markdown, although our initial speculation for its relation

also to shiny turns out to be incorrect.

We can easily extend the above rationale in order to examine closer the relationship

between two or more communities, as long as we limit the investigation to relatively small

communities (remember our 100-node max rule of thumb for visualizations). Say we would

like to see how the “Big Data” and