NodeXL Pro has recently been updated to include support for Twitter API 2.0 and the…
My colleagues on the Facebook Data Team recently posted the results of a study about the diversity of the Facebook user base.
Using surnames from users in the United States and comparing the rates at which those surnames occur in several ethnic populations, they were able to estimate the proportion of Facebook users within different populations over time. The graphic shows a time series of the estimated percentage of U.S. Facebook users from each of 4 races/ethnicities, normalized by the expected percentage. In other words, a value of 100% means the population in question has the same proportional representation on Facebook as it does in the census data. A value over 100% means the population is over-indexed among Facebook users, and a value under 100% means the population is under-represented on Facebook. It appears that Facebook has always been fairly diverse, and apart from over-indexing for Asians/Pacific Islanders the user population is now reasonably representative of the Internet using population in the U.S.
For more details on the study, including the estimation and modeling approaches used, check out the original note on the data team page, or Cameron Marlow’s full repost. Lars Backstrom, Jonathan Chang, Cameron Marlow and Itamar Rosenn conducted this research.