We then study a number of different prominence measures in different networks. For example, h-index based on both the number of citations and the number of papers with high citations is determined by degree (number of distinct co-authors) and the community centrality as papers on more foundational topics are likely to get citation. 

We also look at people with high number of total citations for their top 10 papers. For this measure, the community is not important anymore but the local centrality is. A good paper will get a lot of citations regardless of the community it is in.

How about a different network? For example, for the Internet Movie Database (IMDB) containing actors and movies, we look at three different  measure: movie budget, movie gross and the average rating for the movie in IMDB. Community centrality and global centrality play a role in budget and gross, as movies from central actors and communities get a big investment overall.  This likely results in their overall higher gross as well. Local centrality is not very important, only actors who are universally known appear to contribute to the prominence of their movies. However, for rating, none of the centrality measure play a significant role.

This work appears in the upcoming ASONAM 2013 conference. The PDF can be found here (ALM_Asonam2013.pdf).

Software is available at github.com/rpitrust/prominence/tree/master/CommDistHashmaphttp://asonam.cpsc.ucalgary.ca2_Deconstructing_centrality__how_to_measure_the_prominence_of_individuals_files/AdaliMagdonIsmailLu_cameraready.pdfhttps://github.com/rpitrust/prominence/tree/master/CommDistHashmaphttps://github.com/rpitrust/prominence/tree/master/CommDistHashmaphttp://www.cs.rpi.edu/~sibel/shapeimage_4_link_0shapeimage_4_link_1shapeimage_4_link_2shapeimage_4_link_3