Reading a book by its cover
The ability to help potential advertisers demographically target their messaging is expected to be a big factor in social networking site Facebook's future profitability. So it stands to reason that Facebook has a ‘data team' responsible for mining the mountain of data on the site and extracting useful nuggets from it.
High on the list of any demographic breakdown is race, but Facebook doesn't ask you for your race when you create your profile. So the data team came up with a novel way of estimating the ethnicity of its users - their surnames.
Using data from the US Census Bureau's genealogy project, Facebook was able to see what the ethnic breakdown is for a surname. The three most popular US surnames, and their racial representation are shown below. Api refers to Asia/Pacific, aian - native American, 2prace - two or more races, and hisp - Hispanic.
There were some more statistical techniques employed to allow for the fact that Facebook users may not mirror the broader population, but it comes as no surprise to see that white is by far the biggest racial group.
What Facebook seems keen to prove, however, is that the racial breakdown of its US users is increasingly mirroring that of the broader population - or to be more precise, the Internet population. So we're left with the chart below, in which the proportion of ethnic minorities using Facebook is compared to the proportions for all Internet users.